Upgrading Traefik v1.7 -> v2.1

This week end, I upgraded Traefik from v1.7 to 2.1.

Traefik is used in Spider as edge router as well as load balancer intra cluster.
When I switched from NGINX to Traefik one year ago, there had been a huge improvement in performance and response time stability.

Now, new version is even better! 🙂

Here are the comparison before and after migration over 4h of run:

In short, there is an average improvement of 30% in response times intra cluster, with a slight improvement for edge routers response times.

Do you still wonder if you should change?

Resiliency of Spider system under load :)

Spider endured an unwanted Stress test this morning!
Thanks, it proved it behaved and allowed me to confirm one bug =)
Look what I got from the monitoring:
Load went from 290k/min to 380k/min between 9:00am to 9:20am:
There were so many sessions at once to parsed that the parsing queue what hard to empty:
Thus generating many parsing errors, due to missing packets (short life time)
But all went back to normal afterwards 🙂
CPU and RAM were OK, so I bet the bottleneck was the configuration: limited number of parsers to absorb the load.
But there is a bug: Redis store does not get back to normal completely, some elements are staying in store and not deleted:
That will prevent the system from absorbing too many spikes like this in a row.
Also, some Circuit breakers to Redis opened during the spike (only 4…):
Under this load, the main two Redis instances were enduring 13k requests per second (26k/s combined), with a spike at 31k/s. For only … 13% CPU each. Impressive!
And all this is due to a load test in SPT1 (the 4 whisperers at the bottom)
Conclusion:
  • System is resilient enough 🙂
  • Observability is great and important !!
Bonus: the summary dashboard:
  • A spike at 600k packets/min.
  • Many warnings in logs showing something got bad.
  • CPU on applicative node that went red for some time, and on an ES node as well.
  • Errors on 3 services that may not be scaled enough for this load.
  • And we see the circuit breakers errors due to Redis that suffered.
Really helpful!

Replicas reduction after perfomance optimisations

I had a task that had been hanging around for months:

Reducing replicas and servers count as a consequence of all performance improvements.

Indeed, Spider cluster was still in a replicas configuration built when implementing in Node 8, with generators and using NGINX as gateway.

At this moment, I need many replicas to overcome some NGINX flaws under hi load within Docker.
Except that now… The cluster is handling between 2x and 3x more load, and constantly!

So I did the work of computing theoretical replicas count, and reducing them until it felt reasonable.

Overall results

And the result is great:

  • From 221 cumulated replicas on operational services,
  • It got reduced to… 36 ! 🙂 I removed 185 replicas from the cluster!

Globaly, it is interesting to note that removing replicas:

  • Reduced drastically the RAM usage, which was the goal: – 14GB
  • Decreased alos the total CPU usage.
    • I think this is due to less CPU overhead in changing process

Thanks to the memory reduction, I moved from x7 M5 AWS instances to x5 C5, which are cheaper for better CPU (only 4 GB of RAM).
I may remove one server still, because the average load is 90% CPU and 2 GB of RAM. But I’ll wait some time to be sure.

Detailed step by step

Replica reduction was made step by step, and metrics saved for each step.

  • You may notice that reducing replicas made services faster !! This couls be linked to the removal of context changes…
  • And then, moving from 7 servers to 5 servers increase the latency again. For sure… when a server does almost nothing, it is fastest 😉
    • There was also an intermediate step, not on this table, with 7xC5. But latency almost did not change.

Traefik issue

After reducing replicas, I encountered a strange issue:

There were some sporadic 502 error messages from one service calling another one. But the error message was captured in the caller, the callee did not receive the communications!

And indeed, the issue was in Traefik gateway. The case is not so frequent, but it is due to the big difference of socket timeout between Traefik and Node.
Traefik is at 90s, Node… at 5s. Once Node TTL is configured longer that Traefik’s the 502 disappeared.
Reference: https://github.com/containous/traefik/issues/3237

Automated setup ! :-)

Spider configuration crazyness

I figured out that if I wanted easy adoption of Spider, I had to improve the setup.

Indeed, the way to setup Spider was cumbersome! If the infra (Elasticsearch cluster, Swarm applicative cluster) can be automated with Terraform and Ansible, the cluster configuration involves many assets:

  • Microservices configuration: 34 files
  • File beat & metric beats configuration: 2 files
  • JWT keys
  • SMTP configuration
  • Traefik configuration
  • Elasticsearch indices provisioning (and maintenance)
  • Swarm cluster configuration: 10 files

All those files having parameters with consistent values, links and a lot of potential failure in case of wrong configuration.
Moreover update of configuration on evolutions was really a pain involving GIT history, diff and so on.

So I decided to automate Spider setup and update.

Result

The result is … WoW 🙂 (not Blizzard one, my own :-p)
Here is the ‘minimal’ configuration file to have a running Spider installation, from scratch:

22 lines 🙂
Neat !

How it works

Original setup

Setup instructions are summarized in 4 steps/commands and all configuration is in one file now. From this file, a bunch of scripts:

  • Validate the configuration
  • Take configuration templates and create the environment own configuration from them
  • Provision or update the indices
  • Create the original admin account
  • And start the cluster 🙂

Here are the setup instructions:

Configuration templates are available in a GIT repo that will receives updates for all new features.

Upgrades

  • ‘make update’ will rebase the local repo with the trunk
  • After which a ‘make config db cluster’ will be enough to update current installation.
  • And of course, in case of new required setup values in setup.yml, the user will be told, and the update stopped.

The indices update is automatic and based on metadata added to the indices. It handles parent templates, ILM policies, rolling over indices, rolling up indices and static indices.

  • Rolling over, reindexing and cache purging is all handled programmatically., and automatic. This is so goood =)

 

What do you think? Do you like it?
I did a demo of ideas and execution to Streetsmart DevOps team today… and I think we’ll target the same kind of improvement for Streetsmart! Yeah 🙂

New saving options – TcpSessions and Http parsing resources – leading to a huge optimisation !

New saving options

Due to upgrade to Redis 5.0.7, a bug that I’d been tracking for months has been solved (on Redis 😉 ), and now parsing quality is 100% 🙂 apart from clients issues.
Then I moved to next phase: it is possible not to save parsing resources any more, thus leading to a huge optimisation in Elasticsearch usage…

For this, I recently added two new options in Whisperer configuration:

  • Save Tcpsessions
  • Save Http parsing resources

No need to save those?

As for most cases, at least on Streetsmart, we do not work with Tcp sessions and Http parsing resources, there is no use to save them in Elasticsearch.

  • These resources are indeed only used to track parsing quality and to troubleshoot specific errors.
  • These resources are also saved on Elasticsearch for recovery on errors… but as there are no more errors… it is completely possible to avoid their storage.

I had to change the serialization logic in different places, but after a couple of days, it is now stable and working perfectly 🙂

The whole process of packets parsing is then done in memory, in real time and distributed across many web services in the cluster, using only Redis as a temporary store, moving around 300 000 packets by minute 24h/24, amounting to almost 100MB par minute!

That’s beautiful =)

What’s the result?

  • Less storage cost
  • Less CPU cost for Elasticsearch indexation
  • More availability for Elasticsearch on the UI searches 🙂 You must have noticed less timeouts these days?

On UI, when these resources are not available, their hypermedia links are not shown. And a new warning is displayed when moving to TCP / Packet view:

Currently, Tcp sessions and Packets are only saved on SIT1 and PRM Integration platforms. For… non regression 😉

One drawback though

Those resources where rolled up in Elasticsearch to provide information for the parsing quality line at the bottom of the timeline. :-/

As I got this evolution idea AFTER the quality line implementation, I have nothing to replace it yet.
So I designed a new way to save parsing quality information during the parsing, in a pre-aggregated format. Development will start soon!

It is important for transparency as well as non regression checks.

Upload Whisperers now have their own storage rules, and purge is back :)

Thanks to Leo, we just found out that recent evolution with Index Lifecycle Management had removed the possibility to purge the data manually uploaded to Spider.

I then changed a bit the serialization and storage architecture so that Upload Whisperers have dedicated indices and lifecycles.

Thanks to index templates inheritance and a clever move on the pollers configuration, the changes where pretty limited 🙂

So now:

  • Uploaded data are kept for several weeks.
  • Uploaded data can be purged whenever you don’t need them anymore

Small change, big effects – MTU setting in Docker overlay network

Hi,

Thanks to Pawel who shared me this link: https://github.com/moby/moby/issues/37855, I tried and changed the MTU settings of Spider internal cluster overlay network.

docker network create spider --driver overlay --subnet=10.0.0.0/16 --opt com.docker.network.driver.mtu=8000

It’s only like two words to add to the network settings, but it requires you to restart the full cluster (which I don’t do anymore very often ;)).

And the effet was immediate:

  • Around 50% faster intra cluster exchanges
    • See graphic on the bottom right
    • Other graphs show that load is the same and quality of parsing the same as well.

  • Around 10% less CPU usages on cluster nodes
    • But not that there is no change in service CPU usage: Docker daemon is using less CPU

Man, small change, big effect !!
Why isn’t this documented in Docker docs ?!

Thibaut

 

Btw, in the 15min outage, I had time to:

  • Upgrade Docker, and security packages of all servers
  • Upgrade Elasticsearch, Kibana and Metricbeat 😉

Thanks for compatible versions =)

Parsing quality

I recently added a new feature to Spider timeline:

It now shows the parsing quality of Spider for the selected Whisperer(s).

What do I mean by this?

Explanation

Spider captures packets, then sends them to the server that rebuilds the network communications.
Due to the massive amount of packets received (around 4 000/s on Streetsmart platform, 300 million a day), they are processed in real time with many buffers to absorb the spikes.

However, trouble may get in the way in many places when a spike is too big:

  • On client side:
    • The kernel buffer may get full before pcap process all packets
    • The pcap buffer (network capture) may get full as well
    • Spider server may not be able to absorb the load, and the Whisperer buffer may overflow (by backoff)
    • The same may happen for Tcpsessions tracked on client side… but I never saw this happen.
  • On server side:
    • Under too much load, or infrastructure issue, the parsing may not find all packets (Redis buffer full) and fail to rebuild the communication
    • And there is currently a ‘last’ bug that happens under high load as well that sometimes packets cannot be found.. although they SHOULD be there (I’m trying to track this one down, but this is not easy)

All in all, there are many places where it may go wrong and that could explain that you don’t find the communications you expected!

This information of the goood health or not of the system is available in Spider monitoring, but not so easy to process.

Quality line in timeline

Thus I designed an enhancement of the timeline so that you know when the missing data MAY be a Spider issue.. or if it is indeed due to the system under monitoring 😉

Below X axis, you may see underlined parts in red. They tell you that there have been parsing issues in the period. The redder the worse.
(Timeline is showing PIT1 on november 21st)

You will then find detailed information when hovering on them:

Lots of red !!

Don’t be afraid, the quality is good:

  • The light red is the first threshold. Here, we lost 8 packets out of 2 millions! Looks like there has been a surge of load at this time, and Spider could not absorb.
  • Over the full day, on PIT1, we could not parse 11 communications out of 2,2 Millions. That’s 99.9995% quality!

If failures happens too often, the buffers need to be enlarged, or the server scaled up…

 

So now you know, and when it gets really red, call me for support 😉

 

Hide the line

You may hide this quality line – and avoid making the queries – in the timeline menu.
This settings is saved in your user settings.

Last info

This is PRM integration platform. It is much redder.
Indeed, there are many packets lost by pcap. And this is due to the old Ubuntu version!

Indeed, pcap filters are applied after capture on this Ubuntu version, and not before. Thus, the pcap driver captures all packets of all communication in its buffer before filtering and sending them to the Whisperer.

I enlarged pcap buffer from 20 to 60MB on this server, but it is not enough still.
If Yvan, you are missing too many communications, this can be the cause, and we can enlarge the buffer again. It is easy to do, it is done on the UI 😉