You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* "slugids" -- exactly 22 characters, always a slugid
28
27
*`taskGroupId`
29
28
*`taskId`
30
29
30
+
The following remain separate:
31
+
*`organization`, `repository` (limited to 100 characters in taskcluster-github)
32
+
33
+
The rationale for the choice of 38 is the [task status routing key](https://docs.taskcluster.net/docs/reference/platform/taskcluster-queue/references/events#routing-key) which contains
34
+
35
+
| name | max characters |
36
+
|----------------|----------------|
37
+
| routingKeyKind | 7 |
38
+
| taskId | 22 |
39
+
| runId | 3 |
40
+
| workerGroup | N |
41
+
| workerId | N |
42
+
| provisionerId | N |
43
+
| workerType | N |
44
+
| schedulerId | N |
45
+
| taskGroupId | 22 |
46
+
| reserved | 1 |
47
+
48
+
With a total of 9 dots between each of those components.
49
+
So the total length is `5N + 55 + 9 <= 255`.
50
+
Solving for N, we get 38.
51
+
31
52
This change requires changes to schemas and API declarations in services in the monorepo, as well as to `aws-provisioner` and possibly `ec2-manager`, workers, and client libraries.
32
53
Much of this work has [already been done](https://github.com/taskcluster/taskcluster/pull/110) by @OjaswinM.
33
54
34
55
We already have checks on AMQP routing key lengths, and the increased length does not cause any failures in that code.
0 commit comments