32 comments

[ 5.8 ms ] story [ 111 ms ] thread
This website is completely broken while scrolling. Does anyone else this stretching effect?
The React app that parses content from JSON to HTML (whyyyy?) also injects ~60 dynamic style elements (:facepalm:). Not sure whether to blame the Wix framework their app uses, or the site itself.

You would think that people who write comparisons of time-series analytics tools, and put effort into SEO, would profile and optimize their site. I hope Google buries them for their load time.

Throwing Grafana in the mix seems strange. It is comparing apples and oranges.

Also, contrary to what the article says Grafana does alarms as of v4.0 http://docs.grafana.org/alerting/notifications/

Agreed. Grafana is great as a single-pane-of-glass solution for viewing a variety of stats/logs data, but doesn't actually handle the behind-the-scenes work of collecting/storing them...
Came here to say this. And it's not like the author doesn't get that, since grafana is recommended in the piece to be used as the visualization layer for the other two. I think they just wanted to get all three names in the post for some reason.
Indeed. Prometheus and Grafana are complimentary. In fact PromDash was deprecated in favor of Grafana [1]:

"NOTE: PromDash is deprecated. We recommend Grafana for visualization of Prometheus metrics nowadays, as it has native Prometheus support and is widely adopted and powerful."

This seems like an odd oversight for a "comparison" article.

[1] https://github.com/prometheus-junkyard/promdash

(comment deleted)
"Grafana can only visualize time series and it excels in this task over all other, but neither alarm management nor event tracking are part of its core functionality"

This is false, grafana introduced alerting last year.

Shouldn't Prometheus be compared to InfluxDB?
+1

I was going to say Prometheus vs Graphite vs InfluxDB would be an interesting comparison.

If you just want to compare the database aspects that works. Once you get into other features that gets messier as functionality is split across components in different ways.

For example Prometheus vs InfluxDB is not a great comparison, whereas Prometheus+Alertmanager vs InfluxDB+Kapacitor is.

https://prometheus.io/docs/introduction/comparison/ has such a comparison, though it's a few months out of date.

[Disclaimer: Prometheus developer]

Warning, throws a popup in your face to make you subscribe to their mail spam and plays a sound to make you notice the fake chat having a message in the corner.

Overall this is just a low effort linkbait article to build SEO.

This is definitely a clickbait SEO article but the "chat" isnt fake, it's a service provided by intercom.io (and several other companies) to have live chat on your site for visitors and customers to interact with quickly. Thousands of sites have this now and it's a good feature to have.
Yes. I generally use it to tell whoever is on the other side just how annoying and misleading the chat ping sound was and the blinking of the tab title. Intentionally trying to sound and look like a Facebook, Skype or Gmail notification. Thats definitely not a dark pattern. The sort of company that would use this isn't scammy at all.
The ding for the chat is fake, though. There's nobody on the other end waiting for you to return their engagement.
I had opened it and faked interest, waited for 30 minutes, no reply. There is nothing wrong with providing a chat function but in this case it made a notification sound and displayed a "1", implying a message for me. That is just scummy.
Is this what HN is now? A place for press releases masquerading as low-value analysis?
For a fair comparison, Grafana should have been with combined with InfluxDB, which is sort of the default configuration. I've been using Grafana / InfluxDB for a personal domotica project. Given that I run it on cheap virtualised hardware, it's blazingly fast, extremely easy to use (for devs), intuitive and many client libraries for InfluxDB exist (Scala, Java, Python, R, to name a few).

Couple of restrictions:

- It's best for the relationship: "time", "tag", "scalar". Multi-scalar rendering is possible using a plugin, but it's not designed for it.

- Density plots are not supported yet. Limited support for heatmaps.

- Due to JSON serialisation, server-side compression is definitely necessary. With zopfli, I've seen compression to below 3%. This, however, entails a (limited) CPU load.

- Retention and aggregation in influxdb are possible, but it's best to design this upfront.

> combined with InfluxDB, which is sort of the default configuration

I'd say the "default" or most common configuration is still Graphite + Statsd + Grafana.

These comparisons don't really make sense. I just spent a decent chunk of time evaluating the current state of these tools for implementation in a new product, and seeing a comparison of the prom + grafana toolchain with the comparable Influx DB toolchain would have been way more useful.

If anyone is in the middle of this decision, I went with Prometheus and my early experience has been positive. It seems either would have been a good choice though, and the existence of a decent instrumentation library in language I'm using ended up being the deciding factor.

(comment deleted)
Yeah - all of the "suggestions" are a bit weird. For example they say not use Grafana for OpenStack, to use Gnocchi instead.

Gnocchi supports Grafana as a UI. (as do all of the other tools even mentioned on the page)

This is a weird comparison. Grafana + InfluxDB is quite powerful. No mention of InfluxDB at all.
The author's misunderstanding of role and function of these tools played a big part in their declared winners of each section. I'd say on the whole that this comparison article is useless.
Another case of a simple blog being so horribly overengineered that it throws "Some elements on this page did not load. Refresh your site & try again." in my face after loading 230 (!!) javascripts for over 10 seconds.
"This is where Graphite wins over Grafana. Graphite can store time series obtained from other sources (normally, direct monitoring tools) and provide a query language to obtain the stored data. Again, Grafana can be used with Graphite in order to visualize the data stored on its storage backend."

> Graphite wins over Grafana

> Grafana can be used with Graphite

> wins over

> can be used with

Surprised to see this from a CTO. Shows a lack of technical understanding in time series dbs, visualizations, and alerting​.