Random timeout on request

[Mattious]

Hi,

i've implemented a flow to get release information from a node api by using the package azure-devops-node-api.
I've a strange behaviour. It works perfectly but suddenly i get timeout for the request: connect ETIMEDOUT 13.107.42.18:443

The flow when i get a release status (or also queue a release) is

• Connect to azure devops api:

import * as vm from 'azure-devops-node-api';
 public async azureDevOpsConnectWebApi(): Promise<vm.WebApi> {
        this._authentificationHandler = vm.getPersonalAccessTokenHandler(this.PAT);
        let option = null;
        this._vstsWebApiConnect = new vm.WebApi(this.AZURE_DEV_OPS_URL, this._authentificationHandler, option);
        return this._vstsWebApiConnect;
    }

• get the release status

public async getRelease(projectName: string, releaseDefinitionId: number): Promise<releaseInterface.Release> {
       const _self = this;
       try {
// Connect to azure dev api
           const vstsWebApiConnect: vm.WebApi = await AzureDevOpsHelper.getInstance().azureDevOpsConnectWebApi();
           const releaseApiConnect: releaseApi.IReleaseApi = await vstsWebApiConnect.getReleaseApi();
           const release: releaseInterface.Release = await releaseApiConnect.getRelease(projectName, releaseDefinitionId);
           return release;
       } catch (error) {
           LogUtils.Instance.error(`** AzureDevOpsReleaseHelper - getRelease ${projectName} - ${releaseDefinitionId}: ${error}`);
           throw error;
       }
   }

The time out can occurs on the connection on azure devops api or during get release.

Functionnality can work 50 times but suddenly he next 5 requests failed. Or directly failed and suddenly works.

Can you provide me support or explain what i can do to avoid that ?

Thanks in advance

[damccorm]

This seems like its probably just network issues, I can't get a repro. With that said, one thing you can do to make this moreresilient is to use retries

azure-devops-node-api/api/interfaces/common/VsoBaseInterfaces.ts

Line 93 in cf9a873

maxRetries?: number;

[damccorm]

Note also that you can increase the timeout -

azure-devops-node-api/api/interfaces/common/VsoBaseInterfaces.ts

Line 88 in cf9a873

socketTimeout?: number;

[Metarract]

Also ran into this same issue; I chalked it up to ado throttling me and just increased the timeout to ridiculous amounts but ended up still getting a timeout error and didn't end up seeing any throttling in our usage section for me. Increased it to like 4 hours (assuming the socketTimeout was ms based on testing; if it's in s then it sure was a lot more than that), still timed out before it hit that amount. Set it to 0 hoping it would remove timeouts like some request modules but it still timed out a bunch.

We're trying to get info on around 10k releases for audit purposes, I dunno if the number affects it at all. I assume this is hard to repro since it's extremely intermittent but out of around 10k releases it timed out around 30 times. Also set retries to 5 but going to try with more I guess.

EDIT: Just ran this again with maxRetries of 20 and a super high timeout, call times out regardless within a few minutes once I get to around 6200 api calls, successfully runs a few, times out again. This pattern continues until we hit the end.

[iMPeRiOR]

Hi Guys, I have the same issue on getBranches() method, after a random time it gets timeout, but network is stable, and I've setup high timeout and retries values.

[github-actions]

This issue has had no activity in 90 days. Please comment if it is not actually stale

[rfennell]

I am still seeing this issue intermittently