Platform
Site reliability metrics for the CanadaLogin platform, in the sre schema. Currently incident-response metrics, with other reliability measures to follow.
The relevant tables are:
sre.mttr_current- latest state of every incident, one row per incidentsre.mttr_history- dated snapshots of every incident, preserving how each evolved
These tables come from a hand-maintained spreadsheet; see the catalog Data quality note for caveats. The one that shapes most queries here: the incident timestamps (detected_at, resolved_at, and so on) are stored as America/Toronto strings, not native dates, so a reporting window filters on the YYYY-MM-DD prefix of detected_at rather than a from_date column.
Incident Response Metrics
| Metric | Description |
|---|---|
| Mean Time to Resolution | Average minutes from incident detection to resolution. |
Site Uptime Metrics
Uptime and response-time metrics for the six monitored CanadaLogin sites, in the sre schema. Data comes from the cds-snc/status-statut GitHub repository and is fetched automatically by the uptime-daily-download Lambda at 04:00 ET each day. No manual steps are required.
The relevant tables are:
sre.uptime_summary- one row per site per day; rolling-window uptime percentages and average response timessre.uptime_timeseries- one row per upptime bot commit per site; individual ping events including outage detections
Both tables are partitioned by date=YYYY-MM-DD. Because date is a varchar (derived from the Hive partition path), cast it before doing date arithmetic: CAST(date AS DATE) in SQL; as.Date(date) in R.
| Metric | Description |
|---|---|
| Site Uptime | Rolling-window uptime percentage for each monitored site. |
| Site Response Time | Average HTTP response time for each monitored site, over rolling windows. |
| Outage Events | Individual site outages: when a site went down, when it recovered, and the status code. |