Grafana 2.0 Released | Hacker News Reader

Grafana 2.0 Released

grafana.org

157 pointseloycoto43 comments

One quick note to people writing release notes pages:

Often these pages get posted to news aggregators, and the people who click on those links are not as likely to be familiar with your product. A quick 1-2 line elevator pitch at the top would do wonders.

Couldn't' agree more! I am interested but struggling to get exactly what it does..is it like Gephi..who knows???

Although, I disagree, there is a note on the sidebar.

How does that compare to the ELK (ElasticSearch/LogStash/Kibana) stack? (Is it even right to compare?)

I'm currently evaluating ElasticSearch + Kibana to handle metrics/errors/application logs in our web app, it's impressive but Kibana is lacking a lot of features to be complete IMO.

Good point for ELK in my case is that it runs on windows which is not the case of InfluxDB it seems (running on windows is a requirement in my company unfortunately).

It's a similar web interface (Grafana is based on Kibana), but for different things.

ELK is best for deciphering large amounts of data from logs IMHO. Yes logstash has tons of plugins for third party monitoring systems, but generally speaking you're going to dump your web server access/error logs and syslogs into your ELK setup.

Grafana makes graphite/influxdb more useful and is by far and away the best option available. Usually you're going to be tracking things like cpu/memory/diskio/network traffic, but also anything from StatsD in your Grafana setup.

Combining the two systems into one could be neat, but they work great on their own and have their own benefits.

I'm using Grafana + Influxdb + statsd and I'm quite happy:

- In our django/python apps I'm using statsd client - For server monitoring I'm using diamond - For dashboards I'm using grafana.

I have in a few clicks all company information. I used ELK service, but only for logs but queries were difficult for me.

But, ELK is great for PFsense routers[http://elijahpaul.co.uk/monitoring-pfsense-2-1-logs-using-el...], or you can use http://packetbeat.com/ if you want to monitor all your servers.

Last week I spoke in a local community about my stack. Here you have my slides + examples:

http://acalustra.com/statsd-talk-at-python-vigo-meetup.html

Regards

Thanks for the response. I should probably give a shot to Kibana 3.x because I'm on 4.x and I suspect it's still pretty new and lacking in features because of it.

I could already track system metrics on windows with WMI (well I know CPU and memory is possible at least) and send them to ES but I will need to hand code it to avoid LogStash (too big for this simple job in my case).

ELK is about log analytics, Grafana is about Time series & metric analytics and monitoring. The difference might not seem that big but currently it quite a big difference.

Time series are usually more about being able to collect huge amount of metrics. Metrics that can then be combined, averaged, filtered, put through a processing pipeline (analytical functions), summarized by different intervals. All in order to visualize (usually through graphs) recent live trends or long term trends and statistics.

Grafana is all about maximizing the power and ease of use of the underlying time series store so the user can focus om on building informative and nice looking dashboards. It is also about letting users define generic dashboards through variables that can be used in metric queries, this allows users to reuse the same dashboard for different servers, apps or experiments as long as the metric naming follows a consistent pattern.

Grafana also uses Elasticsearch but not for log analytics, but for annotating graphs with event/log information. At some point in in the coming 1-3 years log analytics and metric analytics & visualization is going to converge and be solved/addressed by the same piece of software. But that is tricky right now without sacrificing either domain.

Monitoring includes:

- fault detection

- metrics

- logs

Grafana + Graphite cover #2. Graphite may also help with #1 if you have something like Sensu watching the metric thresholds.

ELK covers #3.

From the homepage:

> An open source, feature rich metrics dashboard and graph editor for Graphite, InfluxDB & OpenTSDB.

Looks like Prometheus support is in the works too: https://github.com/grafana/grafana-plugins/pull/13

That repo also has Atlas, Zabbix and KairosDB plugins.

I think the most useful feature in Grafana for me has been Templating. It's really powerful and works well in an AutoScaling environment where things come and go. with graphite-web it would be a fight to keep the graphs collecting the right data points, or updating them if your hostname scheme changes, etc.

I use it heavily as well to create generic dashboards for looking at clusters/sites, etc. Its a shame kibana4 nerf'd templating :|

In an autoscaling environment, you just need to keep your metric paths consistent. Then all you need is wildcard queries. E.g., CPU userspace cycles on all your instances should be like:

    provider.region.vpc.group.instance-id.os.cpu.user

E.g.:

    aws.us-west-1.prod.web-front.i-4df567a9.os.cpu.user

Then wildcard queries will provide you aggregate stats at any level you like.

It's really a feature of how Graphite queries work.

--------------------

All that being said, I agree that Grafana is pretty awesome.

That's true. I guess what I prefer about doing it in Grafana is that the interface for managing complex graphs with overlays of different metrics, and ones using wildcards to match groups of metrics is just way easier.

> The Graph panel now supports 3 logarithmic scales, log base 10, log base 32, log base 1024. Logarithmic y-axis scales are very useful when rendering many series of different order of magnitude on the same scale (eg. latency, network traffic, and storage)

Have been waiting eagerly for logarithmic scale support.

What's base 32 used for?

Useful for display series that represent storage space or IO, get get y axis ticks like 1 KB, 32KB, 1MB, 32MB, 1GB http://play.grafana.org/dashboard/db/new-features-in-v20?pan...

I remember wanting to use graphite for a pet project and stumbled across early release Grafana. I'm amazed at all the nagging items I had (even though I still used it) have been resolved. The CORS, shipping with its own backend, not needing elastic search (such a pain!) and sharing graphs are stellar. Truly amazing.

On a very tangentially related note: we've recently open-sourced our .NET port of StatsD/Graphite and it's available under MIT license at https://bitbucket.org/aeroclub-it/statsify

interesting! I am working on a windows build of grafana 2 right now. Does statsify have a /render and /metrics/find api that is compatible with graphite? If it had grafana could use it.

It has something similar to "/render"[0], but it's not Graphite-compatible. There's also support for "/find", but it is currently not the part of the public API.

Adding Graphite-compatible endpoints should be fairly easy and this can really open up a lot of integration opportunities.

[0]: https://bitbucket.org/aeroclub-it/statsify/wiki/Home#markdow...

Could you move that project to github? I think that project has great potential coupled with Grafana.

I will add windows builds to the install and download pages today.

Here is a windows build for grafana 2.0.1 https://s3.amazonaws.com/grafanarel/winbuilds/dist/grafana-2...

What's the recommended backend for this? I tried it with InfluxDB but felt that the feature set wasn't really up there, I couldn't get graphs to display the data I wanted.

We have a graphite instance elsewhere but it's running into disk space issues - will that be an issue here, too?

Graph an is a wholly JavaScript "chart display" front end that hacienda onto a grpahite-query-compatible server. This is usually the graphite server that usually is storing rrd data into a carbon database.

However there are it seems a wide variety of choices. My view is if any form of column orientated data structure is struggling, try sharding your time series long before you swap technologies.

Very nice, congrats on the release! Had it up and running and our Grafana 1.x dashboards imported in about 15min.

I was intrigued reading about the plans for alert support. Any pointers on where to start looking for somebody that may want to help build it? :)

Would love to buy a coffee or beer some time in Stockholm.

That is awesome (that the upgrade was that smooth).

I am going this meetup at DICE on Thursday, https://sites.google.com/site/stockholmgtug/Home/20150423

I wanted to try

http://docs.grafana.org/v2.0/installation/windows/

but when I click the "build from source" link, I get "Sorry, page not found.". :(

strange works now.

Just started using Grafana+Influxdb this week. Could not be happier :)

Congrats.

Thanks!

ElasticSearch was one of the data sources to grafana [1.8.1]. I do not see anything that points to "still works" in the docs. Has this changed?

iframe embedded graphs would've helped me out immensely 8 months ago, So happy to see them now

What is Grafana? I'm having trouble getting a clear explanation of what it is.

Its a dashboard and graph composer for realtime and and historical metric analytic (mostly via rich graphs). Usually used for monitoring or application metrics analytic but can be used in many other scenarios as well.

Picture phpMyAdmin, but for metrics and time series data. You point it to a time series database (such as Graphite or InfluxDB) and the app draws the charts for you with whatever settings you need, and puts them together in pretty dashboards for easy consumption.

Graphite + Grafana are like Cacti or Munin, but orders of magnitude smarter and more powerful.

Grafana is the visualization part of the combo. Graphite is the data storage and query engine.

This looks great!

FYI, the demo site link is broken (missing a : )

Thanks for spotting that, fix in publishing now!

AAwesome! Upgrading this week!

Oh - this is Marvelous. Congrats Torkel!! We plan to put your new multi-tenant features to work!

Now waiting for InfluxDB 0.9 - then we're off to the races.

There's a fork for InfluxDB 0.9: https://github.com/influxdb/grafana/tree/influx-0.9rc4

No idea how well it works, haven't tried it yet, but I would imagine it's still incomplete.

Reply on news.ycombinator.com