consolidate(perf): cut SD writes, log volume, and metrics churn (#486)

* fix(plugins): one bad metrics cache entry should not stop every plugin

Caught live on a rig: every plugin failing, once each, continuously.

    ERROR - src.plugin_system.plugin_manager - plugin geochron operation failed:
    ResourceMetrics.__init__() got an unexpected keyword argument
    'consecutive_failures'

    ERROR - ... plugin text-display operation failed: ...
    ERROR - ... plugin news operation failed: ...
    ERROR - ... plugin odds-ticker operation failed: ...

with /api/v3/health reporting plugin_system: not_initialized while the display
process itself kept running and updating the panel.

`consecutive_failures` is a plugin_health field, not a metrics one.
get_metrics() does ResourceMetrics(**cached), which raises TypeError on a
single unrecognised key, and that exception escapes into plugin_manager and is
reported per plugin. One malformed cache entry takes the whole plugin system
down.

How a health-shaped record came to sit under a plugin_metrics key on that
machine is not established, and I could not finish the diagnosis: the rig went
back into its EIO failure mode partway through -- SSH resetting pre-banner,
systemctl unexecutable -- while the web API kept answering from RAM. Checked
before that: the cache files on disk are correctly shaped and separate, and
CacheManager.get() returns the right record for each key, so it is not a live
key collision. A restored backup mixing two machines' caches is the likeliest
explanation, and that rig had one restored onto it.

Either way the loader should not be brittle enough for the answer to matter.
plugin_health already repairs its records field by field rather than trusting
what is on disk; this does the same. Known fields are kept, unknown ones are
dropped and named once in the log so a genuine schema change stays visible
rather than being silently discarded, and a non-mapping entry no longer raises.

Keeping the known fields matters: discarding the record wholesale would throw
away real call counts and timings because of an unrelated stray key.

Mutation-checked: restoring ResourceMetrics(**cached) fails 6 checks, dropping
the whole record fails the field-preservation check, and dropping unknown
fields silently fails the logging check. 28 tests pass across the resource
monitor and plugin health suites.

* perf(health): stop rewriting a health record on every healthy cycle

Every successful plugin update called record_success(), which persisted the
record unconditionally. In steady state the only fields that had changed were
total_successes and last_success_time -- a counter and a timestamp that
health_monitor surfaces for display and that nothing reads back after a
restart. Nothing alerts on the age of last_successful_update; it is carried in
the metrics dataclass and shown.

Measured on a rig running 24 plugins, all steady-state (0 consecutive
failures, circuit closed): a five-minute sample caught 22 health-file
rewrites, about 4.4 a minute or 6,300 a day. Each write is ~400 bytes through
cache_manager.set(), which writes a file per call, so each one costs a
filesystem block plus an ext4 journal write.

That lands on an SD card, where the unit of cost is an erase-block cycle
rather than the bytes involved, and where wear is what eventually kills the
card. Two cards have already failed on the other rig with the same
signature -- unreadable block device, EIO on exec, sshd unable to read its
host keys.

The circuit breaker still has to survive a restart, so the write is kept for
exactly the fields it is rebuilt from: consecutive_failures, circuit_state,
circuit_opened_time, half_open_start_time. A failure, a circuit opening and a
recovery are all still written the moment they happen. In-memory state is
updated every time either way, so the health API and web UI show what they
always did.

Tested: 100 healthy cycles now perform zero writes after the first, the
counters remain accurate in memory, and a failure, a recovery and a
half-open-to-closed transition each still reach disk. One test kills and
rebuilds the tracker from the cache to prove the breaker's state genuinely
survives what is no longer written.

Mutation-checked both ways: persisting unconditionally again fails the
steady-state test, and widening _DURABLE_FIELDS to include last_success_time
fails it too. The 46 existing health tests pass.

(cherry picked from commit 14abea2d24)
(cherry picked from commit 0f77bd2345)

* perf(vegas): trace the content path at DEBUG instead of INFO

plugin_adapter narrates every step of acquiring content from every plugin --
"Has get_vegas_content", "Native: calling get_vegas_content()", "Native
content returned None", "Has scroll_helper", per-item sizes -- once per plugin
per cycle, all at INFO.

Measured on a live rig: 13,408 log lines an hour, of which 13,366 were INFO
and 35 were WARNING. Roughly 223 lines a minute of string formatting on a Pi
that is also driving the panel, written through journald to the SD card, with
the 35 lines that actually indicate a problem buried among them.

Top repeated messages in that hour:

    717  Scroll progress: elapsed=... total_scrolled=.../... px
    399  [plugin] --> INCLUDED in Vegas scroll
    323  [plugin] content_type=static, display_mode=fixed
    195  [plugin] Has get_vegas_content: True
    195  [plugin] Native: calling get_vegas_content()
    168  [plugin] Native: get_vegas_content() returned None
    168  [plugin] Native content returned None        <- the same fact, twice

54 logger.info calls in plugin_adapter become logger.debug, along with the
per-frame scroll-progress line in scroll_helper. Together those are 3,174 of
the 13,408 lines an hour, a 23% cut, and the ~3,600 odds-manager lines are
addressed separately by ledmatrix-plugins#300.

Nothing is lost: the 19 warning/error/exception calls in the module are
untouched, so real failures still surface at their own level. This is a
logging-level change only -- no control flow, no behaviour.

One INFO call is deliberate and stays. The padding-strip message picks its
level at runtime (`logger.warning if (left and right) else logger.info`) and
test_vegas_plugin_adapter.py pins that choice; it survives because it is not a
direct logger.info call site. That test still passes.

Mutation-checked both ways: reintroducing a single INFO trace fails the guard,
and demoting the warning/error calls along with the trace fails a second guard
written for exactly that mistake. 537 vegas and scroll tests pass.

(cherry picked from commit e496d95dfe)
(cherry picked from commit 8d1e43c15a)

* fix(logging): give the journal the real severity of each line

Everything this process writes to stdout reaches the journal as PRIORITY=6,
whatever the Python level was, because journald has nothing else to go on.
Measured on a live rig over 24 hours:

    lines containing " - ERROR - "      55
    lines containing " - WARNING - "    13
    journald PRIORITY recorded          6, for every one of them

So `journalctl -p err -u ledmatrix` returns nothing while errors are being
logged, and `-p warning` likewise. Triage falls back to grepping message text,
which is slower and unreliable: during this audit a search for "oom" matched
the radar logging "zoom=9" twenty-four times and briefly looked like the OOM
killer had been firing.

systemd reads a leading "<N>" on each stdout line and takes it as the priority
(sd-daemon(3)), so a formatter that prefixes one costs no dependency. Every
line of a multi-line record is tagged, not just the first -- the journal splits
them, and an untagged continuation reverts to the default, which would leave
the body of a traceback filed as informational while its first line was an
error.

Applied only when JOURNAL_STREAM is set, which systemd sets for services whose
output it captures. Run from a terminal, in the emulator or under pytest the
prefixes would be literal noise, and the file handler keeps the plain
formatter for the same reason.

Mutation-checked three ways: prefixing unconditionally fails the
outside-systemd test, prefixing only the first line fails the multi-line test,
and mapping ERROR to 6 fails the level mapping. 39 tests pass across the
logging suites.

(cherry picked from commit 780fca6365)

* fix(logging): let callers see through the journald formatter wrapper

CI caught what local testing could not: two existing tests in
test_logging_config.py assert that setup_logging() selected a
StructuredFormatter or a ContextualFormatter, by checking the console
handler's formatter directly. Wrapping that formatter to tag each line with
its syslog priority makes those assertions false.

They passed locally and failed on the runner because the wrapper is applied
only when JOURNAL_STREAM is set -- absent in a terminal, present in CI. An
environment-dependent break, which is the kind that gets shipped.

The wrapper now exposes the formatter it delegates to, and those two tests
look through it. They are about which formatter format_type selects, and that
behaviour is unchanged; only the object they have to reach for moved.

Verified both ways this time: 39 tests pass with JOURNAL_STREAM set and with
it unset.

* perf(plugins): stop rewriting a plugin's metrics file on every call

Plugin metrics were persisted to the cache inside monitor_call, so every
call by every plugin rewrote a small JSON file. Measured on a running rig:
one plugin's plugin_metrics file changed nine times a minute, with fourteen
such files active. Each is around 350 bytes, which on ext4 costs a 4KB block
plus a journal entry, so the cost is dominated by the write itself rather
than the payload. Cache writes accounted for essentially all of that device's
2.4 MB/min of SD traffic, on a card that wears out and has already failed
twice on the other rig.

Metrics cannot be de-duplicated the way health state can, because call_count
changes on every call and the timings usually do too. So they are rate-limited
instead: at most one write per plugin per 30 seconds.

The in-memory copy stays authoritative and exact -- a plugin's call_count is
still precise the instant after it runs. Only the cross-process snapshot the
web UI reads is delayed, and telemetry up to half a minute old is still a fair
description of a long-running plugin.

reset_metrics clears the throttle timestamp, so a reset is not left showing a
deleted key for the rest of the interval.

Extrapolating the sampled rate, this takes metric writes from roughly 126 a
minute to 28. Health persistence, the other half of the churn, is handled
separately in #475.

Verified by reverting the throttle: the churn test then reports 50 writes for
50 calls. 88 tests pass across resource monitor, plugin system and web API.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01STMbQE4YctTacQXfbYqKuW

* fix: use a monotonic clock and only mark metrics persisted once written

Two review findings on the throttle, both right.

The interval compared wall-clock timestamps. These devices have no RTC, so
the clock jumps by however far off boot-time was the moment NTP first syncs
-- a forward jump would allow an early write, a backward one would stall the
snapshot well past the interval. time.monotonic() is not subject to either.

The timestamp was also recorded before cache_manager.set(). A set() that
raised would buy the next interval's silence without leaving a snapshot
behind, which is the one case where skipping the write is least affordable.
Recorded after the write lands instead, so a failure is retried on the next
call.

Verified by restoring the original ordering: the new test then reports one
write where two are expected.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01STMbQE4YctTacQXfbYqKuW

* Address all six review findings on the perf consolidation

CodeRabbit reported six; this is all six, checked against its own
"Actionable comments posted: 6" rather than against what I happened to
scroll past.

Three are real defects in the code:

1. _under_systemd() trusted the presence of JOURNAL_STREAM.

systemd publishes JOURNAL_STREAM as "dev:ino", and every child process
inherits it -- including one whose stdout has been redirected to a pipe
or a file. The variable outlives the descriptor it describes, so a
subprocess would decide it was talking to the journal and emit the "<N>"
priority prefixes as literal noise into that captured output. That is
exactly the noise the function exists to prevent. It now parses the pair
and fstats stdout, per systemd's own guidance, and returns False for
missing, malformed, mismatched, or unusable descriptors.

2. Cached metrics were not type-checked.

A dataclass does not enforce its annotations, so
ResourceMetrics(call_count="not a number") builds happily and only
fails later, deep inside monitor_call:

    TypeError: can only concatenate str (not "int") to str

Values are now coerced to their declared type at load, where there is
still a cache key to name in the warning, and a value that cannot be
coerced starts the plugin fresh instead of arming a delayed failure.
A numeric string is accepted rather than discarded -- a JSON round-trip
can widen an int, and that is recoverable.

3. The first metrics snapshot was skipped for the first 30s of uptime.

_persist_metrics used 0.0 as the "never written" default. monotonic() is
time since boot on Linux and systemd starts this service at boot, so
`now - 0.0 < 30` was true for the first half-minute of every run: the
throttle swallowed the very first write, the one that matters most after
a restart. The sentinel is now None and the interval is only applied when
a previous write exists.

Three are tests that could pass without testing anything:

4. test_health_write_churn's fake cache stored by reference, so the
   tracker kept mutating the object already in the store -- a record
   could look persisted when no write had happened, which is precisely
   what test_durable_state_survives_a_restart exists to detect. Both
   directions now deep-copy, like a cache that serialises to a file.
   Verified: disabling the one real cache write now fails three tests.

5. test_values_of_the_wrong_type_do_not_raise asserted only that a
   dataclass had been constructed, which was true with the bad value
   still in it. It now asserts the loaded metrics are usable -- the
   field is numeric, and arithmetic on it does not raise -- across four
   kinds of bad value.

6. test_vegas_log_volume counted "logger.error(" in the source text,
   which also matches comments, docstrings and string literals --
   including that module's own docstring, which names those levels. A
   real error call could be demoted with the tally unmoved. It now walks
   the AST, reusing the helper already in the file. Verified: demoting
   all 18 warning/error/exception calls now fails the test.

Verified: every fix mutation-checked by reverting it and confirming the
matching test fails.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01STMbQE4YctTacQXfbYqKuW

---------

Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>

src/common/scroll_helper.py

@@ -328,7 +328,7 @@ class ScrollHelper:
             elapsed_time = current_time - (self.scroll_start_time or current_time)
             # The image already includes display_width padding, so we only need total_scroll_width
             required_total_distance = self.total_scroll_width
-            self.logger.info(
+            self.logger.debug(
                 "Scroll progress: elapsed=%.2fs, target=%.2fs, total_scrolled=%.0f/%d px (%.1f%%)",
                 elapsed_time,
                 self.calculated_duration,

src/logging_config.py

@@ -130,7 +130,12 @@ def setup_logging(
     # Console handler (always add)
     console_handler = logging.StreamHandler(sys.stdout)
     console_handler.setLevel(level)
-    console_handler.setFormatter(formatter)
+    # Under systemd, tag each line so the journal records the real severity
+    # rather than filing everything as informational. The file handler below
+    # keeps the plain formatter: the prefix is meaningful to journald and noise
+    # anywhere else.
+    console_handler.setFormatter(
+        JournalPriorityFormatter(formatter) if _under_systemd() else formatter)
     root_logger.addHandler(console_handler)
     
     # File handler (if specified)
@@ -145,6 +150,84 @@ def setup_logging(
             sys.stderr.write(f"Warning: Could not set up file logging to {log_file}: {e}\n")
 
 
+#: syslog priorities, which is what systemd parses from a "<N>" prefix on
+#: stdout. Mapped from Python's levels.
+_SYSLOG_PRIORITY = {
+    logging.CRITICAL: 2,   # LOG_CRIT
+    logging.ERROR: 3,      # LOG_ERR
+    logging.WARNING: 4,    # LOG_WARNING
+    logging.INFO: 6,       # LOG_INFO
+    logging.DEBUG: 7,      # LOG_DEBUG
+}
+
+
+class JournalPriorityFormatter(logging.Formatter):
+    """Wraps a formatter, prefixing each line with its syslog priority.
+
+    Under systemd everything this process writes to stdout lands in the journal
+    as PRIORITY=6, whatever the Python level was. Measured on a live rig: 55
+    ERROR lines and 13 WARNING lines in a day, every one of them recorded as
+    informational, so `journalctl -p err -u ledmatrix` returned nothing at all
+    while errors were being logged. Anyone triaging has to grep the message
+    text instead, which is both slower and wrong -- a search for "oom" matches
+    the radar logging "zoom=9".
+
+    systemd reads a leading "<N>" on each line and uses it as the priority
+    (sd-daemon(3)), so this needs no extra dependency. Multi-line records get
+    the prefix on every line, since the journal splits them and an unprefixed
+    continuation would fall back to the default.
+    """
+
+    def __init__(self, inner: logging.Formatter):
+        super().__init__()
+        self._inner = inner
+
+    @property
+    def inner(self) -> logging.Formatter:
+        """The formatter doing the actual work.
+
+        Whether journald tagging is applied depends on JOURNAL_STREAM, so it is
+        on under systemd and off in a terminal -- and anything asserting which
+        formatter setup_logging() selected would otherwise get a different
+        answer in CI than on a developer's machine. Exposing the inner one lets
+        those checks stay about format_type, which is what they mean.
+        """
+        return self._inner
+
+    def format(self, record: logging.LogRecord) -> str:
+        text = self._inner.format(record)
+        prefix = f"<{_SYSLOG_PRIORITY.get(record.levelno, 6)}>"
+        return "\n".join(prefix + line for line in text.split("\n"))
+
+
+def _under_systemd() -> bool:
+    """True when stdout really is the journal.
+
+    systemd sets JOURNAL_STREAM to "dev:ino" for services whose output it
+    captures. Presence alone is not enough to act on: the variable is
+    inherited by child processes and survives redirection, so a subprocess
+    whose stdout is a pipe or a file still sees it and would emit the "<N>"
+    priority prefixes as literal noise into that output. systemd's own
+    guidance is to fstat the descriptor and compare st_dev/st_ino, which is
+    what distinguishes "the journal is somewhere in my ancestry" from "my
+    stdout is the journal".
+    """
+    declared = os.environ.get("JOURNAL_STREAM")
+    if not declared:
+        return False
+    try:
+        dev_text, ino_text = declared.split(":", 1)
+        declared_ids = (int(dev_text), int(ino_text))
+    except (ValueError, AttributeError):
+        return False
+    try:
+        stat_result = os.fstat(sys.stdout.fileno())
+    except (OSError, ValueError, AttributeError):
+        # No usable stdout: captured by pytest, detached, or already closed.
+        return False
+    return (stat_result.st_dev, stat_result.st_ino) == declared_ids
+
+
 class PluginLoggerAdapter(logging.LoggerAdapter):
     """LoggerAdapter that stamps every record with its plugin_id.
 

src/plugin_system/plugin_health.py

@@ -178,10 +178,20 @@ class PluginHealthTracker:
             )
         return self._health_state[plugin_id]
     
+    # Fields the circuit breaker is rebuilt from after a restart. Everything
+    # else in a health record is reporting, read only for display.
+    _DURABLE_FIELDS = ('consecutive_failures', 'circuit_state',
+                       'circuit_opened_time', 'half_open_start_time')
+
+    def _durable(self, state: Dict[str, Any]) -> tuple:
+        """The part of a health record whose loss would change behaviour."""
+        return tuple(state.get(field) for field in self._DURABLE_FIELDS)
+
     def record_success(self, plugin_id: str) -> None:
         """Record a successful plugin execution."""
         state = self.get_health_state(plugin_id)
         current_time = time.time()
+        durable_before = self._durable(state)
 
         # Reset consecutive failures
         state['consecutive_failures'] = 0
@@ -199,7 +209,18 @@ class PluginHealthTracker:
             state['circuit_state'] = CircuitState.CLOSED.value
             state['circuit_opened_time'] = None
 
-        self._save_health_state(plugin_id, state)
+        # A healthy plugin reports success every cycle, and in that steady state
+        # the only fields changed above are a counter and a timestamp that
+        # nothing reads back after a restart. Persisting them anyway rewrites a
+        # small file per plugin per cycle: on a rig running 24 plugins, a
+        # five-minute sample measured 22 rewrites, about 4.4 a minute or 6,300 a
+        # day. Those land on an SD card, where the cost is an erase-block cycle
+        # rather than the 400 bytes involved, and where wear is what eventually
+        # kills the card.
+        # In-memory state is still updated every time, so the health API and web
+        # UI show exactly what they did before; only the write is skipped.
+        if self._durable(state) != durable_before:
+            self._save_health_state(plugin_id, state)
 
     def record_failure(self, plugin_id: str, error: Optional[Exception] = None) -> None:
         """Record a failed plugin execution."""

src/plugin_system/resource_monitor.py

@@ -9,7 +9,7 @@ import time
 import logging
 import threading
 from typing import Dict, Optional, Any, Callable
-from dataclasses import dataclass, field
+from dataclasses import dataclass, field, fields
 
 try:
     import psutil
@@ -49,6 +49,20 @@ class ResourceMetrics:
             self.total_execution_time = self.total_execution_time / self.call_count
 
 
+#: How often a plugin's metrics are written to the cache, in seconds.
+#:
+#: Persisting on every call meant a small file rewritten roughly nine times a
+#: minute per plugin. On a rig with fourteen active plugins that was ~126
+#: writes a minute for metrics alone, and since each ~350-byte file costs a
+#: 4KB block plus an ext4 journal entry, it dominated the device's write
+#: volume -- on an SD card, which wears out.
+#:
+#: The in-memory copy stays authoritative and exact; only the cross-process
+#: snapshot the web UI reads is delayed, and telemetry up to half a minute old
+#: is still a fair description of a long-running plugin.
+_METRICS_PERSIST_INTERVAL = 30.0
+
+
 class PluginResourceMonitor:
     """
     Monitors resource usage for plugins.
@@ -75,6 +89,10 @@ class PluginResourceMonitor:
         # Resource metrics per plugin
         self._metrics: Dict[str, ResourceMetrics] = {}
         self._limits: Dict[str, ResourceLimits] = {}
+        # When each plugin's metrics last reached the cache. Metrics change on
+        # every call, so they cannot be de-duplicated the way health state can;
+        # they are rate-limited instead. See _METRICS_PERSIST_INTERVAL.
+        self._metrics_persisted_at: Dict[str, float] = {}
 
         # Thread-local storage for execution tracking
         self._local = threading.local()
@@ -102,6 +120,66 @@ class PluginResourceMonitor:
                 "psutil not available - resource monitoring will be limited to execution time only"
             )
     
+    def _metrics_from_cache(self, plugin_id: str, cached: Any) -> "ResourceMetrics":
+        """Build metrics from a cached record, ignoring anything unrecognised.
+
+        ResourceMetrics(**cached) raises TypeError on a single unexpected key,
+        and that exception escapes into plugin_manager, which reports it as
+        "plugin <id> operation failed". Every plugin fails, and the plugin
+        system never finishes initialising.
+
+        Seen on a live rig: every plugin failing with
+
+            ResourceMetrics.__init__() got an unexpected keyword argument
+            'consecutive_failures'
+
+        which is a plugin_health field, not a metrics one. How a health-shaped
+        record came to sit under a plugin_metrics key on that machine is not
+        established -- a restored backup that mixed two machines' caches is the
+        likeliest explanation -- but the loader should not be brittle enough for
+        it to matter. plugin_health already repairs its records field by field
+        rather than trusting whatever is on disk; this does the same.
+
+        Unknown keys are dropped and named once, so a genuine schema change is
+        visible in the log instead of silently discarded.
+        """
+        if not isinstance(cached, dict):
+            self.logger.warning(
+                "Ignoring cached metrics for %s: expected a mapping, got %s",
+                plugin_id, type(cached).__name__)
+            return ResourceMetrics()
+
+        known = {f.name for f in fields(ResourceMetrics)}
+        unknown = sorted(set(cached) - known)
+        if unknown:
+            self.logger.warning(
+                "Dropping unrecognised field(s) from cached metrics for %s: %s",
+                plugin_id, ", ".join(unknown))
+        # A dataclass does not enforce its annotations, so
+        # ResourceMetrics(call_count="not a number") builds happily and only
+        # blows up later, deep inside monitor_call ("can only concatenate str
+        # (not \"int\") to str"). Coerce here, where there is still a cache
+        # key to name in the warning.
+        declared = {f.name: f.type for f in fields(ResourceMetrics)}
+        usable = {}
+        for key, value in cached.items():
+            if key not in known:
+                continue
+            try:
+                usable[key] = int(value) if declared[key] in ('int', int) else float(value)
+            except (TypeError, ValueError):
+                self.logger.warning(
+                    "Cached metrics for %s have a bad %s (%r); starting fresh",
+                    plugin_id, key, value)
+                return ResourceMetrics()
+        try:
+            return ResourceMetrics(**usable)
+        except (TypeError, ValueError) as e:
+            self.logger.warning(
+                "Cached metrics for %s unusable (%s); starting fresh",
+                plugin_id, e)
+            return ResourceMetrics()
+
     def _get_metrics_key(self, plugin_id: str) -> str:
         """Get cache key for plugin metrics."""
         return f"plugin_metrics:{plugin_id}"
@@ -126,7 +204,7 @@ class PluginResourceMonitor:
                     cache_key, max_age=None, memory_ttl=0 if force_reload else None
                 )
                 if cached:
-                    metrics = ResourceMetrics(**cached)
+                    metrics = self._metrics_from_cache(plugin_id, cached)
                 else:
                     metrics = ResourceMetrics()
                 self._metrics[plugin_id] = metrics
@@ -232,18 +310,8 @@ class PluginResourceMonitor:
                     # CPU is harder to measure per-call, so we track it separately
                     metrics.cpu_percent = self._get_process_cpu_percent()
                 
-                # Persist metrics
+                # Persist metrics, at most once per interval per plugin.
-                cache_key = self._get_metrics_key(plugin_id)
+                self._persist_metrics(plugin_id, metrics)
-                self.cache_manager.set(cache_key, {
-                    'memory_mb': metrics.memory_mb,
-                    'cpu_percent': metrics.cpu_percent,
-                    'execution_time': metrics.execution_time,
-                    'call_count': metrics.call_count,
-                    'total_execution_time': metrics.total_execution_time,
-                    'max_execution_time': metrics.max_execution_time,
-                    'min_execution_time': metrics.min_execution_time if metrics.min_execution_time != float('inf') else 0.0,
-                    'last_update_time': metrics.last_update_time
-                })
             
             # Check limits
             if limits:
@@ -363,6 +431,44 @@ class PluginResourceMonitor:
             summaries[plugin_id] = self.get_metrics_summary(plugin_id)
         return summaries
     
+    def _persist_metrics(self, plugin_id: str, metrics: ResourceMetrics,
+                         force: bool = False) -> None:
+        """Write a plugin's metrics to the cache, at most once per interval.
+
+        Caller must hold ``self._lock``.
+        """
+        # Monotonic, not wall clock: these devices have no RTC, so the clock
+        # jumps by however far off boot-time was the moment NTP first syncs.
+        # A forward jump would allow an early write, a backward one would
+        # stall the snapshot well past the interval.
+        #
+        # The sentinel for "never written" is None, not 0.0. monotonic() is
+        # time since boot on Linux, and systemd starts this service *at* boot,
+        # so `now - 0.0 < 30` was true for the first half-minute of every
+        # single run -- the throttle swallowed the very first snapshot, which
+        # is the one that matters most after a restart.
+        now = time.monotonic()
+        last_written = self._metrics_persisted_at.get(plugin_id)
+        if (not force and last_written is not None
+                and now - last_written < _METRICS_PERSIST_INTERVAL):
+            return
+        cache_key = self._get_metrics_key(plugin_id)
+        self.cache_manager.set(cache_key, {
+            'memory_mb': metrics.memory_mb,
+            'cpu_percent': metrics.cpu_percent,
+            'execution_time': metrics.execution_time,
+            'call_count': metrics.call_count,
+            'total_execution_time': metrics.total_execution_time,
+            'max_execution_time': metrics.max_execution_time,
+            'min_execution_time': (metrics.min_execution_time
+                                   if metrics.min_execution_time != float('inf')
+                                   else 0.0),
+            'last_update_time': metrics.last_update_time,
+        })
+        # Only after the write lands. Marking it first would mean a failed
+        # set() bought the next interval's silence without leaving a snapshot.
+        self._metrics_persisted_at[plugin_id] = now
+
     def reset_metrics(self, plugin_id: str) -> None:
         """Reset metrics for a plugin."""
         with self._lock:
@@ -370,4 +476,7 @@ class PluginResourceMonitor:
                 self._metrics[plugin_id] = ResourceMetrics()
                 cache_key = self._get_metrics_key(plugin_id)
                 self.cache_manager.delete(cache_key)
+                # Let the next call persist immediately rather than leaving the
+                # deleted key absent for the rest of the interval.
+                self._metrics_persisted_at.pop(plugin_id, None)
 

src/vegas_mode/plugin_adapter.py

@@ -83,7 +83,7 @@ class PluginAdapter:
         # into unrelated headlines once the strip refreshed to 9,505px.
         self._offset_shapes: dict = {}
 
-        logger.info(
+        logger.debug(
             "PluginAdapter initialized: display=%dx%d",
             self.display_width, self.display_height
         )
@@ -109,7 +109,7 @@ class PluginAdapter:
         Returns:
             List of PIL Images representing plugin content, or None if no content
         """
-        logger.info(
+        logger.debug(
             "[%s] Getting content (class=%s)",
             plugin_id, plugin.__class__.__name__
         )
@@ -118,7 +118,7 @@ class PluginAdapter:
         cached = self._get_cached(plugin_id)
         if cached is not None:
             total_width = sum(img.width for img in cached)
-            logger.info(
+            logger.debug(
                 "[%s] Using cached content: %d images, %dpx total",
                 plugin_id, len(cached), total_width
             )
@@ -126,46 +126,46 @@ class PluginAdapter:
 
         # Try native Vegas content method first
         has_native = hasattr(plugin, 'get_vegas_content')
-        logger.info("[%s] Has get_vegas_content: %s", plugin_id, has_native)
+        logger.debug("[%s] Has get_vegas_content: %s", plugin_id, has_native)
         if has_native:
             content = self._get_native_content(plugin, plugin_id, offscreen_only)
             if content:
                 total_width = sum(img.width for img in content)
-                logger.info(
+                logger.debug(
                     "[%s] Native content SUCCESS: %d images, %dpx total",
                     plugin_id, len(content), total_width
                 )
                 return self._finalize(content, plugin_id, 'native', plugin)
-            logger.info("[%s] Native content returned None", plugin_id)
+            logger.debug("[%s] Native content returned None", plugin_id)
 
         # Try to get scroll_helper's cached image (for scrolling plugins like stocks/odds)
         has_scroll_helper = hasattr(plugin, 'scroll_helper')
-        logger.info("[%s] Has scroll_helper: %s", plugin_id, has_scroll_helper)
+        logger.debug("[%s] Has scroll_helper: %s", plugin_id, has_scroll_helper)
         content = self._get_scroll_helper_content(plugin, plugin_id, offscreen_only)
         if content:
             total_width = sum(img.width for img in content)
-            logger.info(
+            logger.debug(
                 "[%s] ScrollHelper content SUCCESS: %d images, %dpx total",
                 plugin_id, len(content), total_width
             )
             return self._finalize(content, plugin_id, 'scroll_helper', plugin)
         if has_scroll_helper:
-            logger.info("[%s] ScrollHelper content returned None", plugin_id)
+            logger.debug("[%s] ScrollHelper content returned None", plugin_id)
 
         if offscreen_only:
             # Display capture needs the shared canvas; leave it to the caller.
-            logger.info(
+            logger.debug(
                 "[%s] Needs display capture, deferring to the render thread",
                 plugin_id
             )
             return None
 
         # Fall back to display capture
-        logger.info("[%s] Trying fallback display capture...", plugin_id)
+        logger.debug("[%s] Trying fallback display capture...", plugin_id)
         content = self._capture_display_content(plugin, plugin_id)
         if content:
             total_width = sum(img.width for img in content)
-            logger.info(
+            logger.debug(
                 "[%s] Fallback capture SUCCESS: %d images, %dpx total",
                 plugin_id, len(content), total_width
             )
@@ -226,7 +226,7 @@ class PluginAdapter:
             kept.append(result.image)
 
         if not kept:
-            logger.info(
+            logger.debug(
                 "[%s] All %d image(s) from %s were blank — contributing nothing",
                 plugin_id, len(images), source
             )
@@ -235,14 +235,14 @@ class PluginAdapter:
         trimmed_width = sum(img.width for img in kept)
 
         if trimmed_width < self.config.min_plugin_width:
-            logger.info(
+            logger.debug(
                 "[%s] Trimmed content %dpx is below min_plugin_width %dpx — skipping",
                 plugin_id, trimmed_width, self.config.min_plugin_width
             )
             return None
 
         if trimmed_width != original_width or dropped_blank:
-            logger.info(
+            logger.debug(
                 "[%s] Trimmed %s content: %dpx -> %dpx (%.0f%% reclaimed), "
                 "%d image(s) kept, %d blank dropped",
                 plugin_id, source, original_width, trimmed_width,
@@ -431,7 +431,7 @@ class PluginAdapter:
         """
         if self._offset_shapes.get(plugin_id) != shape:
             if plugin_id in self._item_offsets:
-                logger.info(
+                logger.debug(
                     "[%s] Content is %s now, was %s — restarting the rotation "
                     "rather than resuming at a position that no longer means "
                     "anything", plugin_id, shape,
@@ -579,7 +579,7 @@ class PluginAdapter:
             consumed += 1
 
         if mode == 'truncate':
-            logger.info(
+            logger.debug(
                 "[%s] Width budget %dpx: showing the first %d of %d row(s) "
                 "(%dpx incl. gaps); the rest are not shown (overflow=truncate)",
                 plugin_id, budget, len(selected), len(images), used
@@ -587,7 +587,7 @@ class PluginAdapter:
         else:
             self._record_offset(
                 plugin_id, (start + consumed) % len(images), shape)
-            logger.info(
+            logger.debug(
                 "[%s] Width budget %dpx: showing %d of %d row(s) (%dpx incl. gaps) "
                 "from offset %d; remainder deferred to a later cycle",
                 plugin_id, budget, len(selected), len(images), used, start
@@ -636,7 +636,7 @@ class PluginAdapter:
             if mode != 'truncate':
                 self._record_offset(
                     plugin_id, 0 if end >= img.width else end, shape)
-            logger.info(
+            logger.debug(
                 "[%s] Width budget %dpx: cropped continuous %dpx image to "
                 "[%d:%d] (no item gaps of %dpx+ to align to)%s",
                 plugin_id, budget, img.width, offset, end, min_run,
@@ -674,7 +674,7 @@ class PluginAdapter:
             self._record_offset(
                 plugin_id, 0 if end >= img.width else end_index, shape)
 
-        logger.info(
+        logger.debug(
             "[%s] Width budget %dpx: cropped single %dpx image to [%d:%d] "
             "(%dpx) at item boundaries %d-%d of %d, %s",
             plugin_id, budget, img.width, start, end, end - start,
@@ -698,7 +698,7 @@ class PluginAdapter:
             List of images or None
         """
         try:
-            logger.info("[%s] Native: calling get_vegas_content()", plugin_id)
+            logger.debug("[%s] Native: calling get_vegas_content()", plugin_id)
 
             # Tell the plugin how much width the ticker wants it to use, and
             # narrow the canvas for the duration of the call. A plugin that
@@ -707,7 +707,7 @@ class PluginAdapter:
             # be explicit can read get_vegas_render_width().
             render_width = self.resolve_render_width(plugin, plugin_id)
             if render_width != self.display_width:
-                logger.info(
+                logger.debug(
                     "[%s] Native: requesting %dpx instead of %dpx",
                     plugin_id, render_width, self.display_width
                 )
@@ -735,19 +735,19 @@ class PluginAdapter:
                 plugin._vegas_render_width = None
 
             if result is None:
-                logger.info("[%s] Native: get_vegas_content() returned None", plugin_id)
+                logger.debug("[%s] Native: get_vegas_content() returned None", plugin_id)
                 return None
 
             # Normalize to list
             if isinstance(result, Image.Image):
                 images = [result]
-                logger.info(
+                logger.debug(
                     "[%s] Native: got single Image %dx%d",
                     plugin_id, result.width, result.height
                 )
             elif isinstance(result, (list, tuple)):
                 images = list(result)
-                logger.info(
+                logger.debug(
                     "[%s] Native: got %d items in list/tuple",
                     plugin_id, len(images)
                 )
@@ -768,14 +768,14 @@ class PluginAdapter:
                     )
                     continue
 
-                logger.info(
+                logger.debug(
                     "[%s] Native: item[%d] is %dx%d, mode=%s",
                     plugin_id, i, img.width, img.height, img.mode
                 )
 
                 # Ensure correct height
                 if img.height != self.display_height:
-                    logger.info(
+                    logger.debug(
                         "[%s] Native: resizing item[%d]: %dx%d -> %dx%d",
                         plugin_id, i, img.width, img.height,
                         img.width, self.display_height
@@ -793,13 +793,13 @@ class PluginAdapter:
 
             if valid_images:
                 total_width = sum(img.width for img in valid_images)
-                logger.info(
+                logger.debug(
                     "[%s] Native: SUCCESS - %d images, %dpx total width",
                     plugin_id, len(valid_images), total_width
                 )
                 return valid_images
 
-            logger.info("[%s] Native: no valid images after validation", plugin_id)
+            logger.debug("[%s] Native: no valid images after validation", plugin_id)
             return None
 
         except (AttributeError, TypeError, ValueError, OSError) as e:
@@ -833,20 +833,20 @@ class PluginAdapter:
                 logger.debug("[%s] No scroll_helper attribute", plugin_id)
                 return None
 
-            logger.info(
+            logger.debug(
                 "[%s] Found scroll_helper: %s",
                 plugin_id, type(scroll_helper).__name__
             )
 
             cached_image = getattr(scroll_helper, 'cached_image', None)
             if cached_image is None:
-                logger.info(
+                logger.debug(
                     "[%s] scroll_helper.cached_image is None, triggering content generation",
                     plugin_id
                 )
                 if offscreen_only:
                     # Generating it calls display(), which needs the canvas.
-                    logger.info(
+                    logger.debug(
                         "[%s] scroll_helper cache empty; deferring generation "
                         "to the render thread", plugin_id
                     )
@@ -859,13 +859,13 @@ class PluginAdapter:
                     return None
 
             if not isinstance(cached_image, Image.Image):
-                logger.info(
+                logger.debug(
                     "[%s] scroll_helper.cached_image is not an Image: %s",
                     plugin_id, type(cached_image).__name__
                 )
                 return None
 
-            logger.info(
+            logger.debug(
                 "[%s] scroll_helper.cached_image found: %dx%d, mode=%s",
                 plugin_id, cached_image.width, cached_image.height, cached_image.mode
             )
@@ -888,7 +888,7 @@ class PluginAdapter:
 
             # Ensure correct height
             if img.height != self.display_height:
-                logger.info(
+                logger.debug(
                     "[%s] Resizing scroll_helper content: %dx%d -> %dx%d",
                     plugin_id, img.width, img.height,
                     img.width, self.display_height
@@ -902,7 +902,7 @@ class PluginAdapter:
             if img.mode != 'RGB':
                 img = img.convert('RGB')
 
-            logger.info(
+            logger.debug(
                 "[%s] ScrollHelper content ready: %dx%d",
                 plugin_id, img.width, img.height
             )
@@ -1002,7 +1002,7 @@ class PluginAdapter:
             with self._capture():
                 # Method 1: Try _create_scrolling_display (stocks pattern)
                 if hasattr(plugin, '_create_scrolling_display'):
-                    logger.info(
+                    logger.debug(
                         "[%s] Triggering via _create_scrolling_display()",
                         plugin_id
                     )
@@ -1010,7 +1010,7 @@ class PluginAdapter:
                         plugin._create_scrolling_display()
                         cached_image = getattr(scroll_helper, 'cached_image', None)
                         if cached_image is not None and isinstance(cached_image, Image.Image):
-                            logger.info(
+                            logger.debug(
                                 "[%s] _create_scrolling_display() SUCCESS: %dx%d",
                                 plugin_id, cached_image.width, cached_image.height
                             )
@@ -1022,7 +1022,7 @@ class PluginAdapter:
 
                 # Method 2: Try display(force_clear=True) which typically builds scroll content
                 if hasattr(plugin, 'display'):
-                    logger.info(
+                    logger.debug(
                         "[%s] Triggering via display(force_clear=True)",
                         plugin_id
                     )
@@ -1031,12 +1031,12 @@ class PluginAdapter:
                         plugin.display(force_clear=True)
                         cached_image = getattr(scroll_helper, 'cached_image', None)
                         if cached_image is not None and isinstance(cached_image, Image.Image):
-                            logger.info(
+                            logger.debug(
                                 "[%s] display(force_clear=True) SUCCESS: %dx%d",
                                 plugin_id, cached_image.width, cached_image.height
                             )
                             return cached_image
-                        logger.info(
+                        logger.debug(
                             "[%s] display(force_clear=True) did not populate cached_image",
                             plugin_id
                         )
@@ -1045,7 +1045,7 @@ class PluginAdapter:
                             "[%s] display(force_clear=True) failed", plugin_id
                         )
 
-            logger.info(
+            logger.debug(
                 "[%s] Could not trigger scroll content generation",
                 plugin_id
             )
@@ -1077,15 +1077,15 @@ class PluginAdapter:
         try:
             # Save current display state
             original_image = self.display_manager.image.copy()
-            logger.info("[%s] Fallback: saved original display state", plugin_id)
+            logger.debug("[%s] Fallback: saved original display state", plugin_id)
 
             # Ensure plugin has fresh data before capturing
             has_update_data = hasattr(plugin, 'update_data')
-            logger.info("[%s] Fallback: has update_data=%s", plugin_id, has_update_data)
+            logger.debug("[%s] Fallback: has update_data=%s", plugin_id, has_update_data)
             if has_update_data:
                 try:
                     plugin.update_data()
-                    logger.info("[%s] Fallback: update_data() called", plugin_id)
+                    logger.debug("[%s] Fallback: update_data() called", plugin_id)
                 except (AttributeError, RuntimeError, OSError):
                     logger.exception("[%s] Fallback: update_data() failed", plugin_id)
 
@@ -1097,41 +1097,41 @@ class PluginAdapter:
             # arrangement rather than one that has to be cropped afterwards.
             render_width = self.resolve_render_width(plugin, plugin_id)
             if render_width != self.display_width:
-                logger.info(
+                logger.debug(
                     "[%s] Fallback: rendering at %dpx instead of %dpx",
                     plugin_id, render_width, self.display_width
                 )
 
             with self._capture(), self._render_at(render_width):
                 self.display_manager.clear()
-                logger.info("[%s] Fallback: display cleared, calling display()", plugin_id)
+                logger.debug("[%s] Fallback: display cleared, calling display()", plugin_id)
 
                 # First try without force_clear (some plugins behave better this way)
                 try:
                     plugin.display()
-                    logger.info("[%s] Fallback: display() called successfully", plugin_id)
+                    logger.debug("[%s] Fallback: display() called successfully", plugin_id)
                 except TypeError:
                     # Plugin may require force_clear argument
-                    logger.info("[%s] Fallback: display() failed, trying with force_clear=True", plugin_id)
+                    logger.debug("[%s] Fallback: display() failed, trying with force_clear=True", plugin_id)
                     plugin.display(force_clear=True)
 
                 # Capture the result
                 captured = self.display_manager.image.copy()
 
-            logger.info(
+            logger.debug(
                 "[%s] Fallback: captured frame %dx%d, mode=%s",
                 plugin_id, captured.width, captured.height, captured.mode
             )
 
             # Check if captured image has content (not all black)
             is_blank, bright_ratio = self._is_blank_image(captured, return_ratio=True)
-            logger.info(
+            logger.debug(
                 "[%s] Fallback: brightness check - %.3f%% bright pixels (threshold=0.5%%)",
                 plugin_id, bright_ratio * 100
             )
 
             if is_blank:
-                logger.info(
+                logger.debug(
                     "[%s] Fallback: first capture blank, retrying with force_clear",
                     plugin_id
                 )
@@ -1142,7 +1142,7 @@ class PluginAdapter:
                     captured = self.display_manager.image.copy()
 
                 is_blank, bright_ratio = self._is_blank_image(captured, return_ratio=True)
-                logger.info(
+                logger.debug(
                     "[%s] Fallback: retry brightness - %.3f%% bright pixels",
                     plugin_id, bright_ratio * 100
                 )
@@ -1159,7 +1159,7 @@ class PluginAdapter:
             if captured.mode != 'RGB':
                 captured = captured.convert('RGB')
 
-            logger.info(
+            logger.debug(
                 "[%s] Fallback: SUCCESS - captured %dx%d",
                 plugin_id, captured.width, captured.height
             )

test/test_health_write_churn.py

@@ -0,0 +1,119 @@
+"""A healthy plugin must not rewrite its health record every cycle.
+
+Every successful plugin update called record_success(), which persisted the
+record unconditionally. In steady state the only fields that had changed were
+total_successes and last_success_time -- a counter and a timestamp that
+health_monitor reads for display and that nothing reads back after a restart.
+
+Measured on a rig running 24 plugins: about 17 health-file rewrites a minute,
+roughly 25,000 a day. Each is ~400 bytes, but they land on an SD card where
+the unit of cost is an erase-block cycle, not the byte count, and where wear is
+what eventually kills the card.
+
+The circuit breaker still needs its own state to survive a restart, so the
+write is kept for exactly the fields it is rebuilt from -- and a failure, a
+circuit opening, or a recovery must still be written the moment it happens.
+"""
+import time
+
+import copy
+
+import pytest
+
+from src.plugin_system.plugin_health import PluginHealthTracker, CircuitState
+
+
+class _Cache:
+    """Counts writes; serves back whatever was last written.
+
+    Both directions deep-copy, so this behaves like a real cache that
+    serialises through a file. Storing by reference let the tracker keep
+    mutating the object already in the store, so a record could appear to
+    have been persisted when no write ever happened -- which is precisely
+    what test_durable_state_survives_a_restart is supposed to detect.
+    """
+
+    def __init__(self):
+        self.store = {}
+        self.writes = 0
+
+    def set(self, key, data, ttl=None, **kwargs):
+        self.writes += 1
+        self.store[key] = copy.deepcopy(data)
+
+    def get(self, key, max_age=None, memory_ttl=None, **kwargs):
+        return copy.deepcopy(self.store.get(key))
+
+
+@pytest.fixture
+def tracker():
+    cache = _Cache()
+    t = PluginHealthTracker(cache_manager=cache)
+    return t, cache
+
+
+def test_steady_state_success_stops_writing(tracker):
+    """The regression: 100 healthy cycles used to be 100 SD writes."""
+    t, cache = tracker
+    t.record_success("weather")
+    first = cache.writes
+    for _ in range(100):
+        t.record_success("weather")
+    assert cache.writes == first, (
+        f"{cache.writes - first} redundant writes across 100 healthy cycles"
+    )
+
+
+def test_the_counters_are_still_accurate_in_memory(tracker):
+    """Skipping the write must not skip the bookkeeping."""
+    t, _ = tracker
+    for _ in range(10):
+        t.record_success("weather")
+    state = t.get_health_state("weather")
+    assert state["total_successes"] == 10
+    assert state["last_success_time"] is not None
+    assert state["last_success_time"] <= time.time()
+
+
+def test_a_failure_is_written_immediately(tracker):
+    t, cache = tracker
+    t.record_success("weather")
+    before = cache.writes
+    t.record_failure("weather", RuntimeError("boom"))
+    assert cache.writes > before, "a failure must reach disk"
+
+
+def test_recovery_after_failure_is_written(tracker):
+    """consecutive_failures returning to 0 is durable state changing."""
+    t, cache = tracker
+    t.record_failure("weather", RuntimeError("boom"))
+    before = cache.writes
+    t.record_success("weather")
+    assert cache.writes > before, "recovery must reach disk"
+    assert t.get_health_state("weather")["consecutive_failures"] == 0
+
+
+def test_a_closing_circuit_is_written(tracker):
+    """Success in half-open closes the circuit -- that must survive a restart."""
+    t, cache = tracker
+    state = t.get_health_state("weather")
+    state["circuit_state"] = CircuitState.HALF_OPEN.value
+    state["half_open_start_time"] = time.time()
+    before = cache.writes
+    t.record_success("weather")
+    assert cache.writes > before, "a circuit transition must reach disk"
+    assert t.get_health_state("weather")["circuit_state"] == CircuitState.CLOSED.value
+
+
+def test_durable_state_survives_a_restart(tracker):
+    """What is skipped must genuinely not matter to the breaker."""
+    t, cache = tracker
+    for _ in range(3):
+        t.record_failure("weather", RuntimeError("boom"))
+    for _ in range(50):
+        t.record_success("weather")
+
+    revived = PluginHealthTracker(cache_manager=cache)
+    state = revived.get_health_state("weather")
+    assert state["consecutive_failures"] == 0
+    assert state["circuit_state"] == CircuitState.CLOSED.value

test/test_journald_log_priority.py

@@ -0,0 +1,137 @@
+"""Log lines must reach the journal with their real severity.
+
+Everything this process writes to stdout lands in the journal as PRIORITY=6,
+whatever the Python level was, because journald has no other signal. Measured
+on a live rig over 24 hours: 55 lines containing " - ERROR - " and 13
+containing " - WARNING - ", every one of them recorded as informational. So
+
+    journalctl -p err -u ledmatrix
+
+returned nothing while errors were being logged, and anyone triaging has to
+grep the message text instead. That is slower and it is wrong: a search for
+"oom" also matches the radar logging "zoom=9", which is exactly the false
+positive it produced during this audit.
+
+systemd reads a leading "<N>" on each stdout line and uses it as the priority
+(sd-daemon(3)), so this needs no extra dependency -- and it must only be
+applied when systemd is actually reading, or the prefixes become literal noise
+in a terminal, the emulator, and test output.
+"""
+import logging
+import os
+import sys
+from unittest.mock import patch
+
+import pytest
+
+from src.logging_config import JournalPriorityFormatter, _SYSLOG_PRIORITY, _under_systemd
+
+
+class _Plain(logging.Formatter):
+    def format(self, record):
+        return record.getMessage()
+
+
+def _record(level, msg="hello"):
+    return logging.LogRecord("t", level, "f.py", 1, msg, None, None)
+
+
+@pytest.mark.parametrize("level,expected", [
+    (logging.CRITICAL, 2),
+    (logging.ERROR, 3),
+    (logging.WARNING, 4),
+    (logging.INFO, 6),
+    (logging.DEBUG, 7),
+])
+def test_each_level_maps_to_its_syslog_priority(level, expected):
+    out = JournalPriorityFormatter(_Plain()).format(_record(level))
+    assert out.startswith(f"<{expected}>"), out
+    assert _SYSLOG_PRIORITY[level] == expected
+
+
+def test_error_and_info_are_distinguishable():
+    """The whole point: journalctl -p err must be able to tell them apart."""
+    fmt = JournalPriorityFormatter(_Plain())
+    assert fmt.format(_record(logging.ERROR))[:3] != fmt.format(_record(logging.INFO))[:3]
+
+
+def test_every_line_of_a_multiline_record_is_tagged():
+    """The journal splits them, and an untagged continuation loses its level.
+
+    A traceback is the case that matters -- it is the most important thing in
+    the log and the longest.
+    """
+    out = JournalPriorityFormatter(_Plain()).format(
+        _record(logging.ERROR, "Traceback:\nline one\nline two"))
+    lines = out.split("\n")
+    assert len(lines) == 3
+    assert all(line.startswith("<3>") for line in lines), lines
+
+
+def test_the_message_survives_intact():
+    out = JournalPriorityFormatter(_Plain()).format(_record(logging.WARNING, "disk full"))
+    assert out == "<4>disk full"
+
+
+def test_an_unknown_level_falls_back_to_info():
+    out = JournalPriorityFormatter(_Plain()).format(_record(25))
+    assert out.startswith("<6>")
+
+
+def _stdout_ids():
+    """The dev:ino systemd would publish for this process's stdout."""
+    st = os.fstat(sys.stdout.fileno())
+    return f"{st.st_dev}:{st.st_ino}"
+
+
+def test_prefixing_is_off_outside_systemd():
+    """Otherwise a terminal run, the emulator and pytest all show `<6>`."""
+    with patch.dict(os.environ, {}, clear=True):
+        assert not _under_systemd()
+    with patch.dict(os.environ, {"JOURNAL_STREAM": _stdout_ids()}):
+        assert _under_systemd()
+
+
+def test_an_inherited_journal_stream_does_not_count():
+    """The variable outlives the descriptor it describes.
+
+    systemd sets JOURNAL_STREAM for the service, and every child inherits it
+    -- including one whose stdout has been redirected to a pipe or a file.
+    Trusting the variable alone put literal "<6>" prefixes into that captured
+    output. Only a descriptor whose dev:ino actually matches is the journal.
+    """
+    with patch.dict(os.environ, {"JOURNAL_STREAM": "8:12345"}):
+        assert not _under_systemd(), \
+            "a stale inherited JOURNAL_STREAM was treated as the journal"
+
+
+@pytest.mark.parametrize("value", ["", "not-a-pair", "8", "8:", ":12345",
+                                   "eight:12345", "8:12345:9"])
+def test_a_malformed_journal_stream_is_not_the_journal(value):
+    with patch.dict(os.environ, {"JOURNAL_STREAM": value}):
+        assert not _under_systemd()
+
+
+def test_a_closed_stdout_is_not_the_journal():
+    """os.fstat raises rather than answers; that must not propagate."""
+    with patch.dict(os.environ, {"JOURNAL_STREAM": "8:12345"}), \
+            patch("src.logging_config.sys.stdout") as fake_stdout:
+        fake_stdout.fileno.side_effect = ValueError("I/O operation on closed file")
+        assert not _under_systemd()
+
+
+def test_setup_uses_the_wrapper_only_under_systemd():
+    from src.logging_config import setup_logging
+
+    for env, expect_wrapped in (({}, False),
+                                ({"JOURNAL_STREAM": _stdout_ids()}, True)):
+        with patch.dict(os.environ, env, clear=True):
+            setup_logging()
+            handlers = [h for h in logging.getLogger().handlers
+                        if isinstance(h, logging.StreamHandler)]
+            assert handlers, "no stream handler installed"
+            wrapped = any(isinstance(h.formatter, JournalPriorityFormatter)
+                          for h in handlers)
+            assert wrapped is expect_wrapped, (
+                f"JOURNAL_STREAM={env}: wrapped={wrapped}, expected {expect_wrapped}")
+    logging.getLogger().handlers.clear()

test/test_logging_config.py

@@ -183,15 +183,27 @@ class TestSetupLogging:
         setup_logging()
         assert len(logging.getLogger().handlers) == 1
 
+    @staticmethod
+    def _selected_formatter():
+        """The formatter setup_logging() chose, past any journald wrapper.
+
+        Under systemd the console handler's formatter is wrapped so each line
+        carries its syslog priority. That wrapper is applied only when
+        JOURNAL_STREAM is set, which is true in CI and false in a terminal, so
+        asserting on the handler's formatter directly passes locally and fails
+        on the runner. These tests are about which formatter format_type
+        selects, so they look through the wrapper.
+        """
+        formatter = logging.getLogger().handlers[0].formatter
+        return getattr(formatter, "inner", formatter)
+
     def test_json_format_selects_structured_formatter(self):
         setup_logging(format_type="json")
-        assert isinstance(
+        assert isinstance(self._selected_formatter(), StructuredFormatter)
-            logging.getLogger().handlers[0].formatter, StructuredFormatter)
 
     def test_readable_format_selects_contextual_formatter(self):
         setup_logging(format_type="readable")
-        assert isinstance(
+        assert isinstance(self._selected_formatter(), ContextualFormatter)
-            logging.getLogger().handlers[0].formatter, ContextualFormatter)
 
     def test_log_file_adds_file_handler(self, tmp_path):
         log_file = tmp_path / "test.log"

test/test_metrics_cache_unknown_fields.py

@@ -0,0 +1,128 @@
+"""A malformed metrics cache entry must not take every plugin down with it.
+
+`ResourceMetrics(**cached)` raises TypeError on a single unexpected key, and
+that exception escapes into plugin_manager, which reports it per plugin as
+"plugin <id> operation failed". Every plugin fails and the plugin system never
+finishes initialising -- the health endpoint reports
+`plugin_system: not_initialized` while the display itself keeps running.
+
+Seen on a live rig, once per plugin, continuously:
+
+    ERROR - src.plugin_system.plugin_manager - plugin geochron operation failed:
+    ResourceMetrics.__init__() got an unexpected keyword argument
+    'consecutive_failures'
+
+`consecutive_failures` belongs to plugin_health, not to metrics. How a
+health-shaped record came to sit under a plugin_metrics key on that machine is
+not established -- a restored backup that mixed two machines' caches is the
+likeliest explanation, and the same rig had one restored onto it -- but a
+loader that turns one bad cache entry into a total outage is the part worth
+fixing. plugin_health already repairs its own records field by field rather
+than trusting what is on disk.
+"""
+import logging
+from dataclasses import fields
+from unittest.mock import MagicMock
+
+import pytest
+
+from src.plugin_system.resource_monitor import PluginResourceMonitor, ResourceMetrics
+
+
+class _Cache:
+    def __init__(self, payload=None):
+        self.payload = payload
+
+    def get(self, key, max_age=None, memory_ttl=None, **kwargs):
+        return self.payload
+
+    def set(self, key, data, ttl=None, **kwargs):
+        pass
+
+
+def _monitor(payload):
+    m = PluginResourceMonitor(cache_manager=_Cache(payload))
+    m.logger = logging.getLogger("test")
+    return m
+
+
+#: What the rig actually had under the metrics key.
+HEALTH_SHAPED = {
+    "consecutive_failures": 0, "circuit_state": "closed",
+    "circuit_opened_time": None, "half_open_start_time": None,
+    "last_error": None, "last_failure_time": None,
+    "last_success_time": 1_700_000_000.0, "total_failures": 0,
+    "total_successes": 42,
+}
+
+
+def test_a_health_record_under_the_metrics_key_does_not_raise():
+    """The exact failure: it must degrade, not take the plugin system down."""
+    monitor = _monitor(HEALTH_SHAPED)
+    metrics = monitor.get_metrics(" plugin-a".strip())
+    assert isinstance(metrics, ResourceMetrics)
+
+
+def test_recognised_fields_in_a_mixed_record_are_kept():
+    """Dropping the record wholesale would lose real history unnecessarily."""
+    mixed = dict(HEALTH_SHAPED, call_count=7, memory_mb=12.5)
+    metrics = _monitor(mixed).get_metrics("plugin-b")
+    assert metrics.call_count == 7
+    assert metrics.memory_mb == 12.5
+
+
+def test_a_clean_record_still_loads_unchanged():
+    clean = {f.name: 3 for f in fields(ResourceMetrics)}
+    metrics = _monitor(clean).get_metrics("plugin-c")
+    for name in (f.name for f in fields(ResourceMetrics)):
+        assert getattr(metrics, name) == 3
+
+
+def test_unknown_fields_are_named_in_the_log(caplog):
+    """Silently discarding them would hide a real schema change."""
+    with caplog.at_level(logging.WARNING):
+        _monitor(HEALTH_SHAPED).get_metrics("plugin-d")
+    # getMessage(), not .message: the latter is only populated once a handler
+    # formats the record, so the obvious spelling silently never matches.
+    assert any("consecutive_failures" in r.getMessage() for r in caplog.records), \
+        caplog.text
+
+
+@pytest.mark.parametrize("payload", ["a string", 42, ["a", "list"]])
+def test_a_non_mapping_cache_entry_does_not_raise(payload):
+    metrics = _monitor(payload).get_metrics("plugin-e")
+    assert isinstance(metrics, ResourceMetrics)
+
+
+@pytest.mark.parametrize("bad", [
+    {"call_count": "not a number"},
+    {"memory_mb": None},
+    {"execution_time": {"nested": "junk"}},
+    {"min_execution_time": ["a", "list"]},
+])
+def test_values_of_the_wrong_type_fall_back_to_usable_defaults(bad):
+    """isinstance() alone was not enough.
+
+    A dataclass does not enforce its annotations, so the bad value was simply
+    stored and the old assertion passed -- then monitor_call() raised
+    "can only concatenate str (not \"int\") to str" on the next call. The
+    metrics must come back *usable*, not merely constructed.
+    """
+    monitor = _monitor(bad)
+    metrics = monitor.get_metrics("plugin-f")
+    assert isinstance(metrics, ResourceMetrics)
+
+    field_name = next(iter(bad))
+    assert isinstance(getattr(metrics, field_name), (int, float)), \
+        f"{field_name} came back as {getattr(metrics, field_name)!r}"
+
+    # The real proof: arithmetic on the loaded metrics must not explode.
+    metrics.call_count += 1
+    metrics.total_execution_time += 0.5
+    metrics.update_average_execution_time()
+
+
+def test_a_numeric_string_is_accepted_rather_than_discarded():
+    """JSON round-trips can widen an int to a string; that is recoverable."""
+    metrics = _monitor({"call_count": "7"}).get_metrics("plugin-g")
+    assert metrics.call_count == 7

test/test_resource_monitor.py

@@ -11,7 +11,7 @@ Focus areas:
 import time
 
 import pytest
-from unittest.mock import MagicMock
+from unittest.mock import MagicMock, patch
 
 from src.plugin_system.resource_monitor import (
     PluginResourceMonitor,
@@ -127,3 +127,87 @@ class TestForceReload:
         fresh = mon.get_metrics_summary("p", force_reload=True)
         assert fresh["call_count"] == 7
         assert any(c.kwargs.get("memory_ttl") == 0 for c in cache.get.call_args_list)
+
+
+class TestMetricsPersistenceChurn:
+    """Metrics are telemetry; writing them on every call wore the SD card.
+
+    Each write is a ~350-byte file, which on ext4 costs a 4KB block plus a
+    journal entry. At roughly nine calls a minute per plugin across fourteen
+    plugins it dominated the device's write volume.
+    """
+
+    def test_the_first_snapshot_is_written_even_seconds_after_boot(self):
+        """The throttle must key off "have we written?", not process uptime.
+
+        time.monotonic() is time since boot on Linux, and systemd starts this
+        service at boot. With 0.0 as the missing-timestamp default,
+        `now - 0.0 < 30` was true for the first half-minute of every run, so
+        the very first metrics write -- the one that matters most after a
+        restart -- was silently skipped.
+        """
+        import src.plugin_system.resource_monitor as rm
+        cache = _cache()
+        mon = PluginResourceMonitor(cache, enable_monitoring=False)
+        # 12 seconds after boot: inside the interval, but nothing written yet.
+        with patch.object(rm.time, "monotonic", return_value=12.0):
+            mon.monitor_call("p", lambda: None)
+        writes = [c for c in cache.set.call_args_list
+                  if "plugin_metrics:" in str(c)]
+        assert writes, \
+            "the first snapshot was dropped because the process was young"
+
+    def test_repeated_calls_persist_once_per_interval(self):
+        cache = _cache()
+        mon = PluginResourceMonitor(cache, enable_monitoring=False)
+        for _ in range(50):
+            mon.monitor_call("p", lambda: None)
+        writes = [c for c in cache.set.call_args_list
+                  if c.args and str(c.args[0]).startswith("plugin_metrics:")]
+        assert len(writes) == 1, (
+            f"50 calls produced {len(writes)} metric writes; expected 1")
+
+    def test_the_interval_elapsing_allows_the_next_write(self, monkeypatch):
+        import src.plugin_system.resource_monitor as rm
+        cache = _cache()
+        mon = PluginResourceMonitor(cache, enable_monitoring=False)
+        mon.monitor_call("p", lambda: None)
+        # pretend the interval has passed
+        mon._metrics_persisted_at["p"] -= rm._METRICS_PERSIST_INTERVAL + 1
+        mon.monitor_call("p", lambda: None)
+        writes = [c for c in cache.set.call_args_list
+                  if c.args and str(c.args[0]).startswith("plugin_metrics:")]
+        assert len(writes) == 2
+
+    def test_in_memory_metrics_stay_exact_while_writes_are_skipped(self):
+        mon = PluginResourceMonitor(_cache(), enable_monitoring=False)
+        for _ in range(20):
+            mon.monitor_call("p", lambda: None)
+        assert mon.get_metrics("p").call_count == 20
+
+    def test_reset_lets_the_next_call_persist_immediately(self):
+        cache = _cache()
+        mon = PluginResourceMonitor(cache, enable_monitoring=False)
+        mon.monitor_call("p", lambda: None)
+        mon.reset_metrics("p")
+        mon.monitor_call("p", lambda: None)
+        writes = [c for c in cache.set.call_args_list
+                  if c.args and str(c.args[0]).startswith("plugin_metrics:")]
+        assert len(writes) == 2, "reset should clear the throttle timestamp"
+
+    def test_a_failed_write_does_not_buy_the_next_interval_of_silence(self):
+        """A set() that raises must not count as having persisted.
+
+        Marking the timestamp before the write would leave no snapshot in the
+        cache and still suppress the next 30 seconds of attempts.
+        """
+        cache = _cache()
+        cache.set.side_effect = [OSError("disk full"), None]
+        mon = PluginResourceMonitor(cache, enable_monitoring=False)
+        with pytest.raises(OSError):
+            mon.monitor_call("p", lambda: None)
+        # the very next call must try again rather than skip the interval
+        mon.monitor_call("p", lambda: None)
+        writes = [c for c in cache.set.call_args_list
+                  if c.args and str(c.args[0]).startswith("plugin_metrics:")]
+        assert len(writes) == 2, "a failed write should be retried, not skipped"

test/test_vegas_log_volume.py

@@ -0,0 +1,73 @@
+"""The Vegas content path must trace at DEBUG, not INFO.
+
+plugin_adapter narrates every step of acquiring content from every plugin --
+"Has get_vegas_content", "Native: calling get_vegas_content()", "Native content
+returned None", "Has scroll_helper", the per-item sizes -- and it does that for
+each plugin on each cycle.
+
+Measured on a live rig: 13,408 log lines an hour, of which 13,366 were INFO and
+35 were WARNING. plugin_adapter alone produced 2,457 of them. That is ~223
+lines a minute of string formatting on a Pi that is also driving the panel, all
+of it written through journald to the SD card, and it buries the 35 lines that
+actually indicate a problem.
+
+Nothing is lost by moving it to DEBUG: the 19 warning/error/exception calls in
+the module are untouched, so real failures still surface at their own level.
+
+One INFO call is deliberate and stays -- the padding-strip message chooses its
+level at runtime (`logger.warning if (left and right) else logger.info`) and
+test_vegas_plugin_adapter.py pins it.
+"""
+import ast
+from pathlib import Path
+
+import pytest
+
+ADAPTER = (Path(__file__).resolve().parent.parent / "src" / "vegas_mode"
+           / "plugin_adapter.py")
+
+
+def _logger_calls(path, *levels):
+    """Direct logger.<level>(...) call sites in a module."""
+    tree = ast.parse(path.read_text(encoding="utf-8"))
+    found = []
+    for node in ast.walk(tree):
+        if (isinstance(node, ast.Call)
+                and isinstance(node.func, ast.Attribute)
+                and node.func.attr in levels
+                and getattr(node.func.value, "id", None) == "logger"):
+            found.append(node.lineno)
+    return found
+
+
+def _info_calls(path):
+    """Direct logger.info(...) call sites in a module."""
+    return _logger_calls(path, "info")
+
+
+def test_the_content_path_does_not_trace_at_info():
+    calls = _info_calls(ADAPTER)
+    assert not calls, (
+        "plugin_adapter should trace at DEBUG; found logger.info at lines "
+        f"{calls}. This path runs per plugin per cycle and its output goes to "
+        "the SD card via journald."
+    )
+
+
+def test_real_failures_still_have_a_level_of_their_own():
+    """Demoting the trace must not have swept up the error reporting.
+
+    Counted from the AST rather than with source.count(): the text form also
+    matches comments, docstrings and string literals -- including this
+    module's own docstring, which names these levels -- so a real
+    logger.error() could be demoted while the tally stayed put.
+    """
+    loud = _logger_calls(ADAPTER, "warning", "error", "exception")
+    assert len(loud) >= 15, \
+        f"only {len(loud)} warning/error/exception calls remain: {loud}"
+
+
+def test_the_deliberate_runtime_chosen_level_survives():
+    """The padding-strip message picks its level at runtime; leave it alone."""
+    source = ADAPTER.read_text(encoding="utf-8")
+    assert "logger.warning if (left and right) else logger.info" in source