VPP cli "show run" generates very useful telemetry data describing
runtime VPP node_graph operation and efficiency (vector size, cycles per
packet, ...). Good indicator of VPP node load, health and efficiency.
Useful for automated testing (incl. CSIT), identifying hotspots, trending
what's normal, what's not, and as general operational data for deployments.
Enough arguing for it
We need this data thru VPP API to avoid having to rely on CLI and parsing
plain text output.
Dave Barach suggestion how to approach it here:
> From: csit-dev-bounces@lists.fd.io csit-dev-bounces@lists.fd.io
> On Behalf Of Dave Barach (dbarach)
> Sent: Thursday, May 12, 2016 8:37 AM
> To: Maciek Konstantynowicz (mkonstan) <mkonstan@cisco.com>;
> vpp-dev@lists.fd.io
> Cc: csit-dev@lists.fd.io
> Subject: Re: [csit-dev] [vpp-dev] Propose a way to calculate Reference
> Rate for performance short tests
>
> This would be a simple addition to the vl_api_get_node_graph(...) API.
>
> We shouldn’t try to rush this feature into the 16.06 release, but it’s
> really no big deal...
>
> Thanks… Dave
>
> From: vpp-dev-bounces@lists.fd.io vpp-dev-bounces@lists.fd.io
> On Behalf Of Maciek Konstantynowicz (mkonstan)
> Sent: Thursday, May 12, 2016 8:16 AM
> To: vpp-dev@lists.fd.io Cc: csit-dev@lists.fd.io
> Subject: Re: [vpp-dev] [csit-dev] Propose a way to calculate Reference
> Rate for performance short tests
>
> vpp-dev, Are there any plans for API to get “show run” telemetry, or is
> it already done?
> -Maciek