Conversational Analyst
Observability

System health

Every answer above is only as good as the data underneath. This is the operations view of the entire pipeline — source freshness, data quality runs, and recent incidents.

Pipeline uptime (30d)
99.94%
vs 99.5% SLA
DQ checks passing
142 / 143
last 24h
Avg query latency
438 ms
p95 1.2s
Open incidents
0
1 resolved this week

Source freshness

Real-time replication lag for every connected system

SourceDomainLast updateReplication lagSLAStatus
AcumaticaERP2 min ago< 5 min5 minhealthy
Ring CentralTelephony5 min ago5 min10 minhealthy
SentaDispatch8 min ago8 min15 minhealthy
CRMSales11 min ago11 min30 minhealthy
Snowflake goldWarehouse3 min ago3 min10 minhealthy
WorkdayHR / payroll47 min ago47 min60 minwarning

Data quality

Suites that ran in the last 6 hours. Failed checks block downstream answers automatically.

Revenue reconciliation
18/18 passed·last run 06:12·21s
Service order integrity
24/24 passed·last run 06:12·33s
Customer master de-dup
9/9 passed·last run 06:12·12s
Dispatch → invoice match
13/14 passed·last run 06:12·28s
GL → subledger tie-out
31/31 passed·last run 06:12·1m 04s

Recent incidents

Auto-mitigated where possible. All incidents notify the data team in Slack.

Ring → Customer match rate dipped to 94%resolved
minor · Yesterday, 14:22

6% of inbound Ring calls were not auto-matched to a customer. Likely cause: new spoofed-caller-ID rule. Auto-mitigated; running normally.