Health check

mc-dm:health-check runs eight named probes against the addon's runtime and reports pass / warn / fail per probe. Probe names are stable, output is line-oriented, and the command exits non-zero when any probe fails — feed it straight to a monitoring script.

php cmd.php mc-dm:health-check

Sample output:

PASS  addon_active                  active
PASS  storage_mount_writable        round-trip OK
PASS  db_schema_sync                27 xf_mc_dm_* tables present
PASS  db_columns_sync               7 entity selects ok
PASS  job_queue_depth               0 addon jobs queued
PASS  cron_schedule                 7 active crons scheduled
PASS  registry_loadable             4/4 registries loaded
PASS  option_sanity                 2 required options present

Probes

Name	What it checks	warn condition	fail condition
addon_active	The addon is installed and not disabled	—	xf_addon row missing or active = 0
storage_mount_writable	A file written through internal-data://mc-downloads/ can be read back and deleted	—	Write threw, or the round-trip read returned nothing
db_schema_sync	All addon tables exist	—	Fewer than 14 xf_mc_dm_* tables present
db_columns_sync	Each entity selects without a column-not-found error	—	At least one entity throws on SELECT
job_queue_depth	Backlog of addon-owned jobs in xf_job	More than 0 jobs queued	—
cron_schedule	Addon cron entries are active and not stuck	A cron's next_run is more than 7 days in the past	—
registry_loadable	All four extension registries (Gate, Processor, Source, Stats) instantiate cleanly	—	Any registry throws during construction
option_sanity	Required options resolve to a value	—	Any required option is missing or empty

Exit codes

• 0 — every probe passed (or warned).
• 1 — at least one probe reported fail.

warn does not affect the exit code. A backlog of jobs or a slightly stale cron is not necessarily an outage; it is information for the operator.

Wiring into monitoring

Cron the command at whatever interval matches your alerting tolerance and pipe the output to your log shipper:

*/5 * * * * cd /var/www/forum && php cmd.php mc-dm:health-check >> /var/log/mc-dm-health.log 2>&1

For pull-based monitoring (Nagios, Sensu, Zabbix), shell out to the command and treat the exit code as the check state. The line-oriented output is parseable with awk '$1 == "FAIL"' if you want per-probe alerting downstream.