Andreas Öhlund comments

Results 90 comments of


                                            Andreas Öhlund

CLA Assistant Status Occasionally Not Updated

Just tried to deliver the webhook and it succeeded again, this time is worked though and the CLA check is now green. Are there any server-side logs we can look...

Defining message contracts is unnecessarily complicated

Just a quick status update since as of core 7.2 we can now receive messages with no conventions in place I did a quick spike this morning and supporting a...

The end of processing a message is captured before the message is acked in the queueing system and is causing the audit headers for processing and critical time to be incorrect

Forked off https://github.com/Particular/NServiceBus/issues/3494. Labeled as core v6

The end of processing a message is captured before the message is acked in the queueing system and is causing the audit headers for processing and critical time to be incorrect

After triage today we decided to remove the v6 launch label since this is not a must fix for release. Description updated

The end of processing a message is captured before the message is acked in the queueing system and is causing the audit headers for processing and critical time to be incorrect

Reopening, I had put "fix #thisissue" in the pull request text by misstake

The end of processing a message is captured before the message is acked in the queueing system and is causing the audit headers for processing and critical time to be incorrect

Yes we can do it in a minor (the pull needs to be reverted back to have the perfcounters on by default) But this means that we go live with...

The end of processing a message is captured before the message is acked in the queueing system and is causing the audit headers for processing and critical time to be incorrect

> can we consider not including these counters in v6 and adding them in a minor release or even a separate package? Are you saying to ship v6 without the...

The end of processing a message is captured before the message is acked in the queueing system and is causing the audit headers for processing and critical time to be incorrect

> How about shipping with not so accurate counters? That is a possibility, the snag (I guess we could potentially live with it) is that the accuracy of the perfcounters...

The end of processing a message is captured before the message is acked in the queueing system and is causing the audit headers for processing and critical time to be incorrect

Talked to @danielmarbach and @boblangley and we decided to not include this in v6 since: - The inaccuracy is only really significant for broker transports where latency is an issue...

The end of processing a message is captured before the message is acked in the queueing system and is causing the audit headers for processing and critical time to be incorrect

Since the performance counters are driven by the metrics package I have raised https://github.com/Particular/NServiceBus.Metrics/issues/174 to track the fix there and refocused this issue on the message headers that we include...