Closed
Bug 1141254
Opened 10 years ago
Closed 10 years ago
node client: Implement socket level timeouts and pooling
Categories
(Taskcluster :: General, defect)
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: jlal, Assigned: jlal)
References
Details
Node has a number of underlying capabilities to change the behavior of http requests in error conditions and in cases with many requests (connection pools/keep alive).
We use superagent which by default turns off the use of agents (https://github.com/visionmedia/superagent/blob/608cda651db3ac1559a604032878bae3ac16efd7/lib/node/index.js#L233) and has no timeouts. This can cause some issues when attempting to run many requests (you end up with unbounded number of open connections which then take forever to timeout...).
The intent here is to fix this by:
- adding a default timeout of 30s for all client http requests
- use an agent (https://nodejs.org/api/http.html#http_class_http_agent) of the correct http/https type (based on base url) with a default number of allowed open sockets (10 per domain?)
Assignee | ||
Updated•10 years ago
|
Assignee: nobody → jlal
Status: NEW → ASSIGNED
Assignee | ||
Comment 1•10 years ago
|
||
landed here https://github.com/taskcluster/taskcluster-client/commit/cb1a2d381cec3871aa34cd41ebeaebaa5f705c11
Status: ASSIGNED → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Updated•10 years ago
|
Component: TaskCluster → General
Product: Testing → Taskcluster
Target Milestone: --- → mozilla41
Version: unspecified → Trunk
Comment 2•10 years ago
|
||
Resetting Version and Target Milestone that accidentally got changed...
Target Milestone: mozilla41 → ---
Version: Trunk → unspecified
You need to log in
before you can comment on or make changes to this bug.
Description
•