BIG-IP Report

Perhaps it's a SOAP/REST difference?

Good guess! It is. SOAP got more permissions with a read-only role than REST does.

It's a read only account. I'm only using one traffic group on that F5.

Could you please send me the results from this REST endpoint from both devices?

curl -sku admin:password https://<F5-management>/mgmt/tm/cm/failover-status

Please go ahead and clean the output from sensitive information (if any) before posting. 🙂

It's a read only account. I'm only using one traffic group on that F5. I should mentioned that I'm using an older version, on 5.3.1 on a Windows Server and do not have these two issues there and I use the same account. Perhaps it's a SOAP/REST difference?

Howdy!

Certificate information missing

Is the user that BigIP report is using auditor or above? Else it won't be able to read the certificate information. 

Pair missing data

Sounds like the collector believes both devices are passive in which case it won't collect any data by default. Are you using multiple traffic groups on these devices by any chance?

Hi Patrik, I'm using the Kubernetes version 5.6.4 (also tried 5.6.2) and running into a couple of issues:

• The certificates pane is empty - this error is in the log - Error loading certificates. {"code":400,"message":"\"recursive\" unexpected argument","errorStack":[],"apiError":26214401} (line 1126)
• A certain pair of F5s do not display any data in the webpage. There isn't any error in the data collector log. It oddly will not display their hostname and also says both of them are not active.

I'm getting a invalid invite code from your Discord link.

These devices are about to be decomissioned in the upcoming months so no way we can put money and effort into installing new servers there. I was thinking about some simple mitigation instead.

Oh darn. I recognize the challenges with the Chinese firewall.

Lived in Wuhan for a year (before Covid) and there were constant challenges with getting outside the famous wall. Afraid that if it is the national firewall that gives you trouble there's not that much to do at the moment except for running a separate BigIPReport within the borders of China.

Since all objects are arrays you could even run a simple script to concatenate the data from the Chinese BigIPReport to the main report. I think this would be a bit more robust that relying on a few hundred API calls through the firewall and frankly very easy to do.

Well, we see error e-mails couple of times a day.

One of my colleagues is aware of the auth issue, we did some upgrades recently so this seems to be the root cause.

But indexing errors we see pretty much every day at least once, normally the same devices. Config is not big but connection might be interrupted cause they are sitting behind the Great Chinese Firewall and we get all sorts of crap from it 🙂 Is there any way to influence timeouts and retry-periods, eventually on a per-host basis?

Hi there!

• Does it happen consistently?
• Have you read this part of the FAQ?
  https://loadbalancing.se/bigipreport-rest/#Unable_to_get_a_token_from_the_device
• I can also recommend this one (authentication part):
  https://loadbalancing.se/2021/03/28/installing-troubleshooting-and-running-bigip-ingress-controller/#Troubleshooting

If it's inconsistent it could also be poor connection/management provisioning being too small or the config too large. We're on discord if you want to discuss it/get help:
https://discord.gg/RzmjgneW

Kind regards,
Patrik

Go ahead, we have a pre-prod server running LB-monitor as well. Marius is an ex-colleague of ours 🙂

We are having issues with the script though:

bigip-ext-abc-1.domain.com:Failed to get auth token
bigip-ext-bde-5.domain.com:Failed to get auth token
bigip-cloud-01.domain.com:Failed to get auth token
bigip-ext-bde-1.domain.com does not seem to have been indexed
bigip-ext-abv-5.domain.com does not seem to have been indexed
bigip-ext-abv.domain.com does not seem to have been indexed

These errors are raised is on a daily basis. Probably due to a temporary connection glitch/slow response. Any timeouts we can adjust to alleviate the problem?

Hi there!
There's a pull request from Marius Bauer which includes support for policies. Both Tim and I are old school iRulers (yeah yeah, I know that policies are faster) so we would like to know if there's any users out there willing to test Marius branch?

If there is I'll release a beta package for testing.

Kind regards,
Patrik