Memory footprint increased after host instance changed

[pregress]

Is your question related to a specific version? If so, please specify:

4.1036.2.2 & 4.1038.400.1

What language does your question apply to? (e.g. C#, JavaScript, Java, All)

C#, dotnet 9

Question

Is it possible the memory footprint of the host process changed with an additional 50 MB ?
This would no be a problem if it's just start up memory, but we notice there is no drop.

On multiple (linux)functions we see that the avg working set memory is increased after a change of the host instance. While the amount of requests/function executions is stable, and there was no new code deployed.

Sample graph:

We also saw the following properties where added by "system" at the same time:
clientCertExclusionEndPoints
inboundIpv6Address
outboundIpv6Addresses
possibleInboundIpv6Addresses
possibleOutboundIpv6Addresses
sandboxType
useQueryStringAffinity

Removed:
defaultHostNameScope

I create this here instead of azure support so this is publicly available.

[kshyju]

Can you share what Linux SKU is this?

[pregress]

Can you share what Linux SKU is this?

We have it on the following SKU: B1, P0V3, P1v3

[kshyju]

We also saw the following properties where added by "system" at the same time:
clientCertExclusionEndPoints
inboundIpv6Address
outboundIpv6Addresses
possibleInboundIpv6Addresses
possibleOutboundIpv6Addresses
sandboxType
useQueryStringAffinity

Could you clarify where you are seeing this?

[pregress]

We also saw the following properties where added by "system" at the same time:
clientCertExclusionEndPoints
inboundIpv6Address
outboundIpv6Addresses
possibleInboundIpv6Addresses
possibleOutboundIpv6Addresses
sandboxType
useQueryStringAffinity

Could you clarify where you are seeing this?

Using Change Analysis (Preview) or Function => Diagnose and Solve problems => Availability & Performance => Application changes

[pregress]

@kshyju any root cause for this?

[kshyju]

@pregress Apologies for the delayed response. I conducted some initial investigations but did not observe an increase in memory footprint between these two versions. My testing was performed in an isolated functions runtime environment, not on the exact SKU. I will dedicate more time next week to delve deeper into this issue.

Are you still seeing the metric remain elevated?

[pregress]

@pregress Apologies for the delayed response. I conducted some initial investigations but did not observe an increase in memory footprint between these two versions. My testing was performed in an isolated functions runtime environment, not on the exact SKU. I will dedicate more time next week to delve deeper into this issue.

Are you still seeing the metric remain elevated?

Yes, memory usage is still higher, then before.

[pregress]

@kshyju a week has passed, any update?