


I see a strange error when changing my triggers/rules using whisk alarms/interval. Example: I had a fully running action (interval 5 mins) and wanted to change the interval. after "aio app undeploy" I changed just the "interval" and deployed again, running into issues (see screenshot). then I played around with names of triggers/rules and interval - suddenly running again. There is no obvious spelling mistake in the manifest.yaml so I think there must be another issue. see the list here:
Looking closer at the entries I can see that it tryes to run actions with "nodejs:10" (failing) sometimes changeing to "nodejs:12" (success). in the manifest.yaml the action in question has "runtime: 'nodejs:12'" as parameter, maybe that is the issue? but why does it change to "nodejs:10"? any option to force using only "nodejs:12"?
Here are the ID's of the last 5 entries in case somebody wants to have a closer look:
success: 3211537210944900915372109449000d
success: b929585cefc74d43a9585cefc72d4336
appError: bd59cc225e43436799cc225e43236714
appError: df38dc8dc7b14e18b8dc8dc7b13e18e1
success: b30e8a7b238f47218e8a7b238f772104
success: 87aa869a058b406faa869a058b406fbc
remark: why are there always 2 entries for each interval? especially confused about the "undefined" entry...
Views
Replies
Sign in to like this content
Total Likes
Hi Urs - you could try to get the list of triggers and rules being deployed in your namespace, as well as getting details of each of them for troubleshooting. Getting list with `aio rt trigger list` and details with `aio rt trigger get triggerID`. Same syntax for `rule`.
For each interval, the first entry is the alarm trigger itself (which gives undefined, because it's not an action), and the second is the real invocation being triggered.
Views
Replies
Sign in to like this content
Total Likes