How to decide what to probe:
Rank your critical user journeys. List which systems and interactions each one covers. Skip any that aren't adding new coverage.
Maximum structural validation, minimal redundancy.
https://t.co/4JjS7NGxUf
Getting an org to invest in probers is hard. Deciding what to probe is harder.
Most solve it by probing everything — expensive fast, and it ends with a fleet nobody trusts.
Probe systemic issues and critical user journeys. Not every feature.
https://t.co/jPgAFq60b6
1/ Self-reviewing LLM-generated code matters. Here's a real example from a Cloudprober PR where Claude Code's solution worked but had some reliability problems:
https://t.co/qvzfnZ0Nrl
So glad to see @cloudprober being used in more and more places. This is from @rubykaigi (a popular Ruby conference in Japan).
See https://t.co/yPZJCQAruC for the known list of large organization using Cloudprober.
. @cloudprober v0.14.1 matures browser probe functionality further with better UI, filtering, retries, and cleanup; rolls out a convenient single-run mode with JSON output; and includes improvements across networking, Prometheus integration, documentation, and build reliability.
https://t.co/uU4BVH6e7A
Cloudflare’s Scaling with Safety blog (link in reply) highlights @cloudprober as key to verifying performance during their migration! 🛠️ Built with grit at Google Cloud, it’s rewarding to see its trust in action. #Cloudprober#SRE#monitoring
_ @cloudprober
news: Added TLS support to TCP probe: https://t.co/OrAi3iM41l… Add TLS check to TCP probe by just adding `tls_handshake` to `tcp_probe`.
Fixed a really old bug in Cloudprober's UDP probe today:
https://t.co/aaE2FyQyr3
Our UDP probe is special - it allows timeout to be bigger than interval, and uses multiple src ports to increase the network coverage. As a result its implementation is relatively more complex.
Since it happened much after the release, it's not very well known that Cloudprober supports alerting directly. Some organizations are already making use of this feature. I wanted to make sure more people are aware of it 🙂
https://t.co/TKAsE8CosU
Working on adding "browser" probe type to @Cloudprober: https://t.co/34tgxuyTWE. It's coming along pretty nicely. This was one of the longest known gaps in #cloudprober, compared to large synthetic #monitoring providers like DataDog.
We're using #playwright to run the actual tests.
A monitoring software that makes it super-easy to monitor availability and performance of various components of your system
#golang#SysOps
https://t.co/3UEIsGm2wT
Why do we need probers? I've answered this question so many times in the last 2-2.5 years, I decided to write a blog post about it.
https://t.co/Ox2UZMdO8y
#cloudprober#reliability#sre
And it's done. Cloudprober's logger uses log/slog now. Logs are now in logfmt format (key-value pair) by default, and can be switched to JSON format with the "--logfmt=json" flag: https://t.co/2OiGaTugAd
#golang#opensource.
Cloudprober website has a new look: https://t.co/tHtcWjA8Kb. I've also updated the #Kubernetes documentation[1] to introduce the the newly added k8s_targets, and simplified the targets documentation in general.
[1]- https://t.co/ymrqLOXL5n
New release of Cloudprober (https://t.co/raB0E06GE4). Minor bug fixes and improvements. In case you missed the last release announcement, Cloudprober now has a shiny probes status UI at <prober-host>:9313/probestatus (https://t.co/1xet1WSK9M)
MacOS's recvmsg() behavior tripped Cloudprober up. It works differently for IPv4 and IPv6, but the mitigation that we used for that, applying a darwin-specific socket option - IP_STRIPHDR, stopped working for some reason (or maybe it never worked). https://t.co/bfXYaIX5Lu.