What does processing health actually tell me? #83
|
How should I interpret project enrichment counts versus global processor state? Does a clean health response prove source maps or scheduled processing work? |
Replies: 1 comment
|
Use MCP Global Event statuses need their own interpretation. A false Processing uses best-effort background requests plus hourly recovery, not a durable queue. Current ordinary OTLP events generally lack the normalized debug-ID linkage needed for mapping; a release label alone does not supply it. |
Use MCP
get_processing_healthor authenticatedGET /api/:project_id/processing. The response separates the selected project's enrichment and map counts fromprocessing_state, whose scope is global across the installation.Global
last_started_at,last_finished_at,processed,failed, andhas_errordescribe the recorded processor run. Counts are attempts in that run, not lifetime totals or the selected project's throughput. A null state means no run is recorded; an idle invocation can leave the recorded health unchanged.Event statuses need their own interpretation.
not_applicablemeans no usable source-map reference applies.pendingcan mean a missing map, unchecked work, or resumable budge…