Pinot Latency Monitor ¶

The pinot-based APDEX features are the next most common after the CLP-based ones. Apache Pinot

We have a monitor that will check different tables in our Apache Pinot data store. You can view the Pinot panel in our APDEX Data Latency dashboard.

Alerting Definition ¶

1. What the alert looks like ¶

A sample Pinot alert is pretty simple.

The main thing to look for is the Pinot table that was being queried. In this example it is the callanalyzer_report_metrics table.

Note: The 'JFK02' is NOT the Pinot data source but really references the STAP ESaaS that this query result came from in the alert. You can ignore that.

2. Check Pinot ¶

To check on Pinot you should go to the Staging WAP Grafana.

https://wapgrafana-staging.webex.com/explore?orgId=1

One recommendation is to go to an APDEX DAG for Pinot and get a query from the logs to use, and then adjust the time as needed.

SELECT 
    DATETIMECONVERT("dateTimeStamp", '1:MILLISECONDS:EPOCH', '1:MILLISECONDS:EPOCH', '60:SECONDS') as "timestamp",
     locusDataCenter as "shard" ,
    DISTINCTCOUNT(identifiers_correlationId) as "satisfied"
FROM
    callanalyzer_result_metrics
WHERE
     "dateTimeStamp" >= 1707175800000 AND "dateTimeStamp" < 1707176700000  AND resultName = 'pstn.result' AND isConvergedArchitectureEnabled = 'true'AND callType IN ('PSTN_ONLY_DIALIN', 'PSTN_DIALIN') AND outcome = 'success'
GROUP BY
    "timestamp",shard
ORDER BY 
    "timestamp",satisfied
Limit 50000

If you see a gap in data you can ask about this in the Ask ROMA teams space.