Heavy Report Definition affect on Robotics Queue
Hi All,
I’m writing this post since we recently faced an issue with robotics queue and looking for an explanation form anyone who has an idea or faced the same issue.
We had two 7.3.1 environments (let’s say “Dev” and “Test”). Requirement was to fetch some data from some third-party websites via bots.
Case type referring here (let’s say “Provider”) has 7 stages and the robotic automation happens conditionally on stage 5.
In stage 3, we are fetching provider data through a data page where its source was a report definition. This happens for all cases.
Always there will be one result to check the provider type. In Dev environment, we had 50+ records in the data types browsed by the report definition. Report definition has joins with three different data types.
In dev environment, everything works as expected including the bot execution. Also in Test, cases were processed as expected initially. But, after months, work assignments which got assigned to robotics queue were not picked by bots. Also, the server memory usage went up to 14-16GB and eventually crashed and server (we are running on a WebSphere server) When such occurs, restart was required to bring back the server.
After analyzing, the only difference we found is the record count in two of the tables which were browsed by the report definition used by the data page mentioned above. Test environment had over 140K records.
Hi All,
I’m writing this post since we recently faced an issue with robotics queue and looking for an explanation form anyone who has an idea or faced the same issue.
We had two 7.3.1 environments (let’s say “Dev” and “Test”). Requirement was to fetch some data from some third-party websites via bots.
Case type referring here (let’s say “Provider”) has 7 stages and the robotic automation happens conditionally on stage 5.
In stage 3, we are fetching provider data through a data page where its source was a report definition. This happens for all cases.
Always there will be one result to check the provider type. In Dev environment, we had 50+ records in the data types browsed by the report definition. Report definition has joins with three different data types.
In dev environment, everything works as expected including the bot execution. Also in Test, cases were processed as expected initially. But, after months, work assignments which got assigned to robotics queue were not picked by bots. Also, the server memory usage went up to 14-16GB and eventually crashed and server (we are running on a WebSphere server) When such occurs, restart was required to bring back the server.
After analyzing, the only difference we found is the record count in two of the tables which were browsed by the report definition used by the data page mentioned above. Test environment had over 140K records.
We imported the 140K records into Dev and then the same issue was reproduced in Dev. We did a full trace and there were no errors found. Rollbacking the records bought back the system into normal state.
Fix we did :
Initially we renamed the data page and it solved the issue for a while. We did this because we thought this was caused by a caching issue. Then we replaced the data page’s source report definition with an activity to fetch data through a Connect-SQL and now both Dev and Test works well with 140K records.
Issue is fixed and still I can’t understand why this is affecting bots and crashing the server only when stage 7 is executed. If anyone has an idea, that will be appreciated. Hope above explains the issue since I can’t publish the exact data due to information security.
Thanks in Advance.