The monitorRepo folder is where all monitoring details and metrics are stored in your Integration Server. Over time, this folder’s size may fluctuate—growing as new data is collected and shrinking when older data is purged.
If you need to reduce the size of the monitorRepo directory, follow the steps below to adjust retention settings and clean up unnecessary files.
Steps to Reduce monitorRepo Size
1. Stop the Integration Server
Before making changes, ensure the Integration Server is stopped to prevent any conflicts.
2. Adjust Retention Days in monitor.cnf
Navigate to:
<installPath>/IntegrationServer/instances/<instanceName>/config/monitor/monitor.cnf
Modify the retentionDays parameter to a smaller value (e.g., between 1 and 5 days).
- A lower value means less historical data is retained, reducing storage usage.
3. Clear the Index Folder
Run the following command to remove old index files:
rm -rf <installPath>/IntegrationServer/instances/default/<instanceName>/monitorRepo/core/index
4. Adjust Data Deletion Batch Size (Optional)
In IS extended settings, modify:
watt.server.statisticsDataCollector.deletionInterval.batchSizeInHour=<hours>
- The default value is 12, but reducing it to 1 ensures smaller, more frequent purges.
5. Restart the Integration Server
Apply the changes by restarting the server.
6. Manually Trigger Data Deletion (Optional)
If you don’t want to restart the server, you can manually invoke the data deletion service:
http://<host>:<port>/invoke/wm.server.datacollector:deleteData
Other useful services:
- wm.server.datacollector:stopEngine
- wm.server.datacollector:restartEngine
- wm.server.datacollector:startEngine
- wm.server.datacollector:status
Monitoring Database Space Usage
To estimate required database space, run the following queries:
Row Counts in Key Monitoring Tables
sql
SELECT COUNT(*) FROM IS_MONITOR;
SELECT COUNT(*) FROM IS_SERVICE_STATS;
SELECT COUNT(*) FROM IS_SERVER_DETAILS;
SELECT COUNT(*) FROM IS_SERVER_STATS;
Unique Services and Invocation Counts
sql
SELECT NAMESPACE, COUNT(NAMESPACE)
FROM IS_SERVICE_STATS
GROUP BY NAMESPACE
ORDER BY NAMESPACE;
IS_MONITOR Row Counts by Timestamp
sql
-- For MS SQL / Similar Databases
SELECT TS AS TS_msec,
CONVERT(DATETIME, TS/86400000.0) + '1970' AS Timestamp,
COUNT(TS) AS RowCounts
FROM IS_MONITOR
GROUP BY TS
ORDER BY TS DESC;
-- For Oracle / Other Databases
SELECT TS AS TS_msec,
TO_DATE('19700101', 'YYYYMMDD') + (1/24/60/60/1000) * TS AS Timestamp,
COUNT(TS) AS RowCounts
FROM IS_MONITOR
GROUP BY TS
ORDER BY TS DESC;
Aggregated Records by Day and Category
sql
SELECT COUNT(TS) AS NUM_ROWS,
CATEGORY,
DATEADD(d, TS / (1000 * 60 * 60 * 24), '19700101') AS TS_DATE
FROM IS_MONITOR
GROUP BY DATEADD(d, TS / (1000 * 60 * 60 * 24), '19700101'), CATEGORY
ORDER BY TS_DATE DESC;
Conclusion
By adjusting retentionDays, clearing old index files, and optimizing purge intervals, you can effectively manage the monitorRepo folder size. Regularly monitoring database tables ensures efficient storage usage while retaining necessary metrics.