At cloud scale, telemetry and health monitoring is very critical. We should define a framework that allows platform owners to add their own telemetry hooks. Telemetry service should be designed to make this data accessible and store in resilient way (like blackbox during plane crash).

Is there any workgroup that drives this feature “Platform telemetry and health monitoring” ?

Wishlist

BMC telemetry :

Linux subsystem

Uptime
CPU Load average
Memory info
Storage usage ( RW )
Dmesg
Syslog
FDs of critical processes
Alignment traps
WDT excursions

IPMI subsystem

Request and Response logging par interface with timestamps ( KCS, LAN, USB)
Request and Response of IPMB

i. Request , Response, No of Retries

Misc

Critical Temperature Excursions

i. Minimum Reading of Sensor

ii. Max Reading of a sensor

iii. Count of state transition

iv. Retry Count

Count of assertions/deassertions of GPIO and ability to capture the state
timestamp of last assertion/deassertion of GPIO

Thanks

~Neeraj

From: openbmc <openbmc-bounces+neladk=microsoft.com@lists.ozlabs.org> On Behalf Of vishwa
Sent: Wednesday, May 8, 2019 1:11 AM
To: Kun Yi <kunyi@google.com>; OpenBMC Maillist <openbmc@lists.ozlabs.org>
Subject: Re: BMC health metrics (again!)

Hello Kun,

Thanks for initiating it. I liked the /proc parsing. On the IPMI thing, is it only targeted to IPMI -or- a generic BMC-Host communication kink ?

Some of the things in my wish-list are:

1/. Flash wear and tear detection and the threshold to be a config option
2/. Any SoC specific health checks ( If that is exposed )
3/. Mechanism to detect spurious interrupts on any HW link
4/. Some kind of check to see if there will be any I2C lock to a given end device
5/. Ability to detect errors on HW links

On the watchdog(8) area, I was just thinking these:

How about having some kind of BMC_health D-Bus properties -or- a compile time feed, whose values can be fed into a configuration file than watchdog using the default /etc/watchdog.conf always. If the properties are coming from a D-Bus, then we could either append to /etc/watchdog.conf -or- treat those values only as the config file that can be given to watchdog.
The systemd service files to be setup accordingly.

We have seen instances where we get an error that is indicating no resources available. Those could be file descriptors / socket descriptors etc. A way to plug this into watchdog as part of test binary that checks for this ? We could hook a repair-binary to take the action.

Another thing that I was looking at hooking into watchdog is the test to see the file system usage as defined by the policy.
Policy could mention the file system mounts and also the threshold.

For example, /tmp , /root etc.. We could again hook a repair binary to do some cleanup if needed

If we see the list is growing with these custom requirements, then probably does not make sense to pollute the watchdog(2) but
have these consumed into the app instead ?

!! Vishwa !!

On 4/9/19 9:55 PM, Kun Yi wrote:

Hello there,

This topic has been brought up several times on the mailing list and offline, but in general seems we as a community didn't reach a consensus on what things would be the most valuable to monitor, and how to monitor them. While it seems a general purposed monitoring infrastructure for OpenBMC is a hard problem, I have some simple ideas that I hope can provide immediate and direct benefits.

1. Monitoring host IPMI link reliability (host side)

The essentials I want are "IPMI commands sent" and "IPMI commands succeeded" counts over time. More metrics like response time would be helpful as well. The issue to address here: when some IPMI sensor readings are flaky, it would be really helpful to tell from IPMI command stats to determine whether it is a hardware issue, or IPMI issue. Moreover, it would be a very useful regression test metric for rolling out new BMC software.

Looking at the host IPMI side, there is some metrics exposed through /proc/ipmi/0/si_stats if ipmi_si driver is used, but I haven't dug into whether it contains information mapping to the interrupts. Time to read the source code I guess.

Another idea would be to instrument caller libraries like the interfaces in ipmitool, though I feel that approach is harder due to fragmentation of IPMI libraries.

2. Read and expose core BMC performance metrics from procfs

This is straightforward: have a smallish daemon (or bmc-state-manager) read,parse, and process procfs and put values on D-Bus. Core metrics I'm interested in getting through this way: load average, memory, disk used/available, net stats... The values can then simply be exported as IPMI sensors or Redfish resource properties.

A nice byproduct of this effort would be a procfs parsing library. Since different platforms would probably have different monitoring requirements and procfs output format has no standard, I'm thinking the user would just provide a configuration file containing list of (procfs path, property regex, D-Bus property name), and the compile-time generated code to provide an object for each property.

All of this is merely thoughts and nothing concrete. With that said, it would be really great if you could provide some feedback such as "I want this, but I really need that feature", or let me know it's all implemented already :)

If this seems valuable, after gathering more feedback of feature requirements, I'm going to turn them into design docs and upload for review.

--

Regards,

Kun