Thanks for sharing telemetry and for sharing the reload experience. I think that we have some data to go on in trying to replicate this.
Some thoughts:
- After some variable time the user sees “Error Loading Tasks”
Problem onset is inconsistent. Once it happens the problem occurs consistently forever.
It seems that the device triggers getting tasks for all contacts:
"rules-engine:tasks:all-contacts": {
"max": 20787268,
"min": 4306,
"sum": 20791574,
"count": 2,
"sumsqr": 432110529445460
},
"tasks:load": {
"max": 20818586,
"min": 20818586,
"sum": 20818586,
"count": 1,
"sumsqr": 433413523039396
},
According to these entries, generating tasks does succeed, and tasks are loaded successfully once - recording telemetry happens after task generation.
"rules-engine:tasks:dirty-contacts": {
"max": 1727,
"min": 0,
"sum": 29359,
"count": 20,
"sumsqr": 50702993
},
Having 1727 dirty contacts makes me think that emissions have automatically expired at 7 days post calculation.
Another valuable metric for attempting replicate this would be the total number of tasks that you expect this user to generate, and I don’t think we have information about that in telemetry. Does anyone happen to know this number?
- Anecdotally this affects about 200 users per month
Is there anything common between these users? A common role? or a common workflow?