Episodes
Events from different systems that look like one story. Grouped by the platform, confirmed by you.
An episode is a proposal, not a finding
Grouping is based on the signals listed against each one. Confirming an episode is a human decision, and it is recorded as yours.
Open episodes4
Highest confidence first. Confidence describes the grouping, not the diagnosis.
All services
| Episode | Span | Correlated on | Contents | Confidence | State | |
|---|---|---|---|---|---|---|
EP-0042 Daraja token service exhaustion, recurring Daraja API, M-PESA API Gateway | 14 Aug to 1 Sep, 18 days | Shared configuration itemSignature match | 11 incidents, 1 changes, 2 threads, 1 documents | 86% | Unconfirmed | |
EP-0041 Prepaid billing batch overrun, Tuesdays Prepaid Billing | 5 Aug to 31 Aug, 4 occurrences | Signature matchTime proximity | 4 incidents, 1 changes, 1 threads | 74% | Unconfirmed | |
EP-0039 USSD certificate rotation fallout USSD Gateway | 21 Aug, same evening | Linked recordTime proximity | 2 incidents, 1 changes, 1 threads, 1 documents | 91% | Confirmed | |
EP-0038 Fuliza limit check latency Fuliza | 12 Aug to 30 Aug, intermittent | Time proximity | 3 incidents, 0 changes | 52% | Weak |
EP-0042, what the platform grouped
Daraja token service exhaustion, recurring
ChangeDaraja API gateway node patch, first attempt
CHG0011088ServiceNow, 14 Aug 22:10
Incident504 on token endpoint, node 3
INC0047812ServiceNow, 15 Aug 09:12
ConversationDrain node 3 before restarting the token service
TEAMS-noc-paymentsTeams, 15 Aug 09:41
Incident504 on token endpoint, node 3
INC0047994ServiceNow, 22 Aug 08:55
DocumentToken expiry is not monitored end to end
IA-2025-114Uploaded, 12 Aug
ProblemRecurring gateway connection pool exhaustion
PRB0003318ServiceNow, 23 Aug 11:02
Incident504 on token endpoint, node 3
INC0048231ServiceNow, 1 Sep 12:18
Why these are grouped
Shared configuration item
CI-000418, Daraja API
Signature match
504 on /oauth/v1/generate
Recurrence
11 occurrences in 18 days
Time proximity was not used here. The occurrences are weeks apart, so it would have grouped the wrong things.
What happens on confirm
Confirming records your decision in Smart Ops. It does not update the problem record; that stays a write into ServiceNow, and it needs Phase 3.