Placeholder ticket to track adding metrics. More to be added here later — this is just the first one captured while reviewing #294.
Webhook: failed-update visibility
launcher/fly/src/main/kotlin/me/madhead/tyzenhaus/launcher/fly/routes/Webhook.kt deliberately always responds 200 OK and catches all exceptions (to avoid Telegram retry storms on poison updates). That's an intentional, reasonable choice — but it makes processing failures invisible outside the logs: Telegram won't retry, and there's no signal that anything broke.
Action: emit an error/failure metric in the catch block (e.g. a updates.failed counter), so persistent failures are observable. Micrometer / MetricsRepository is already wired up, so this is cheap.
Optionally, also count total updates handled / ignored (0-match in the pipeline) for a success/failure ratio.
Other metrics
TODO — to be filled in later (more metrics planned).
Placeholder ticket to track adding metrics. More to be added here later — this is just the first one captured while reviewing #294.
Webhook: failed-update visibility
launcher/fly/src/main/kotlin/me/madhead/tyzenhaus/launcher/fly/routes/Webhook.ktdeliberately always responds200 OKand catches all exceptions (to avoid Telegram retry storms on poison updates). That's an intentional, reasonable choice — but it makes processing failures invisible outside the logs: Telegram won't retry, and there's no signal that anything broke.Action: emit an error/failure metric in the
catchblock (e.g. aupdates.failedcounter), so persistent failures are observable. Micrometer /MetricsRepositoryis already wired up, so this is cheap.Optionally, also count total updates handled / ignored (0-match in the pipeline) for a success/failure ratio.
Other metrics
TODO — to be filled in later (more metrics planned).