In our last Chaos day we experimented with the Camunda Exporter MVP. After our MVP we continued with Iteration 2, where we migrated the Archiver deployments and added a new Migration component (allows us to harmonize indices).

Additionally, some fixes and improvements have been done to the realistic benchmarks that should allow us to better compare the general performance with a realistic good performing benchmark.

Actually, this is what we want to explore and experiment with today.

• Does the Camunda Exporter (since the last benchmark) impact performance of the overall system?
  • If so how?
• How can we potentially mitigate this?

TL;DR; Today's, results showed that enabling the Camunda Exporter causes a 25% processing throughput drop. We identified the CPU as a bottleneck. It seems to be mitigated by either adjusting the CPU requests or removing the ES exporter. With these results, we are equipped to make further investigations and decisions.

After a long pause, I come back with an interesting topic to share and experiment with. Right now we are re-architecture Camunda 8. One important part (which I'm contributing to) is to get rid of Webapps Importer/Archivers and move data aggregation closer to the engine (inside a Zeebe Exporter).

Today, I want to experiment with the first increment/iteration of our so-called MVP. The MVP targets green field installations where you simply deploy Camunda (with a new Camunda Exporter enabled) without Importers.

TL;DR; All our experiments were successful. The MVP is a success, and we are looking forward to further improvements and additions. Next stop Iteration 2: Adding Archiving historic data and preparing for data migration (and polishing MVP).

Camunda Exporter​

The Camunda Exporter project deserves a complete own blog post, here is just a short summary.

Our current Camunda architecture looks something like this (simplified).

It has certain challenges, like:

• Space: duplication of data in ES
• Maintenance: duplication of importer and archiver logic
• Performance: Round trip (delay) of data visible to the user
• Complexity: installation and operational complexity (we need separate pods to deploy)
• Scalability: The Importer is not scalable in the same way as Zeebe or brokers (and workload) are.

These challenges we obviously wanted to overcome and the plan (as mentioned earlier) is to get rid of the need of separate importers and archivers (and in general to have separate application; but this is a different topic).

The plan for this project looks something like this:

We plan to:

1. Harmonize the existing indices stored in Elasticsearch/Opensearch
   • Space: Reduce the unnecessary data duplication
2. Move importer and archiver logic into a new Camunda exporter
   • Performance: This should allow us to reduce one additional hop (as we don't need to use ES/OS as a queue)
   • Maintenance: Indices and business logic is maintained in one place
   • Scalability: With this approach, we can scale with partitions, as Camunda Exporters are executed for each partition separately (soon partition scaling will be introduced)
   • Complexity: The Camunda Exporter will be built-in and shipped with Zeebe/Camunda 8. No additional pod/application is needed.

Note: Optimize is right now out of scope (due to time), but will later be part of this as well.

MVP​

After we know what we want to achieve what is the Minimum viable product (MVP)?

We have divided the Camunda Exporter in 3-4 iterations. You can see and read more about this here.

The first iteration contains the MVP (the first breakthrough). Providing the Camunda Exporter with the basic functionality ported from the Operate and Tasklist importers, writing into harmonized indices.

The MVP is targeting green field installations (clean installations) of Camunda 8 with Camunda Exporter without running the old Importer (no data migration yet),

Our first goal of this experiment is to use a benchmarks to derive new optimized cluster configuration that can handle at least 100 tasks per second, while maintaining low backpressure and low latency.

For our experiment, we use a newly defined realistic benchmark (with a more complex process model). More about this in a separate blog post.

The second goal is to scale out optimized cluster configuration resources linearly and see if the performance scales accordingly.

TL;DR;

We used a realistic benchmark to derive a new cluster configuration based on previous requirements.

When we scale this base configuration linearly we see that the performance increases almost linearly as well, while maintaining low backpressure and low latency.

Chaos Experiment​

Expected​

We expect that we can find a cluster configuration that can handle at 100 tasks second to be significantly reduced in resources in relation to our smaller clusters (G3-S HA Plan) since these can process significantly above our initial target.

We also expect that we can scale this base configuration linearly, and that the processing tasks rate to grow initially a bit faster than linearly due to the lower relative overhead, and if we keep expanding further to flatten due to the partition count being a bottleneck.

Actual​

Minimal Requirements for our Cluster​

Based on known customer usage, and our own previous experiments, we determined that the new cluster would need to create and complete a baseline of 100 tasks per second, or about 8.6 million tasks per day.

Other metrics that we want to preserve and keep track are the backpressure to preserve user experience, guarantee that exporting speed can keep up with the processing speed, write-to-import latency which tells us how long it takes for a record to be written to being imported by our other apps such as the operator.

Reverse Engineering the Cluster Configuration​

For our new configurations the only resources that we are going to change are the ones relevant to the factors described above. These are the resources allocated to our zeebe-brokers, gateway and elasticSearch.

Our starting point in resources was the configuration for our G3-S HA Plan as this already had the capability to significantly outperform the current goal of 100 tasks per second.

The next step was to deploy our realistic benchmark, with a payload of 5 customer disputes per instance and start 7 instances per second, this generated approximately 120 tasks per second (some buffer was added to guarantee performance).

After this we reduced the resources iteratively until we saw any increase in backpressure, given that no there was no backlog of records, and no significant increase in the write to import latency.

The results for our new cluster are specified bellow in the tables, where our starting cluster configuration is the G3-S HA Plan and the new configuration cluster is the G3 - BasePackage HA.

G3-S HA	CPU Limit	Memory Limit in GB
operate	2	2
operate.elasticsearch	6	6
optimize	2	2
tasklist	2	2
zeebe.broker	2.88	12
zeebe.gateway	0.9	0.8
TOTAL	15.78	24.8

G3 - BasePackage HA	CPU Limit	Memory Limit in GB
operate	1	1
operate.elasticsearch	3	4.5
optimize	1	1.6
tasklist	1	1
zeebe.broker	1.5	4.5
zeebe.gateway	0.6	1
TOTAL	8.1	13.6

Reduction in Resources for our Optimized Cluster​

	CPU Reduction (%)	Memory Reduction (%)
zeebe.broker	47.92	62.5
zeebe.gateway	33.33	-25.0
operate.elasticsearch	50.00	25.0

Total cluster reduction:

	G3-S HA	G3 - BasePackage HA	Reduction (%)
CPU Limits	15.78	8.1	49
Memory Limits	24.8	13.6	45

The process of reducing the hardware requirements was donne initially by scaling down the resources of the zeebe-broker, gateway and elasticSearch. The other components were left untouched, as they had no impact in our key metrics, and were scaled down later in separate experiences to maintain user experience.

Scaling out the Cluster​

Now for the scaling procedure we intend to see if we can linearly increase the allocated resources and having a corresponding performance increase, while keeping the backpressure low, low latency, and user experience.

For this we started with the G3 - BasePackage HA configuration and incremented the load again until we saw any increase in backpressure, capture our key metrics and repeated the process for the cluster configuration resources respectively multiplied by 2x, 3x, and 4x.

This means that the resources allocated for our clusters were:

	Base 1x	Base 2x	Base 3x	Base 4x
CPU Limits	8.7	17.4	26.1	34.8
Memory Limits	14.9	29.8	44.7	59.6

The results in the table bellow show the performance of our several cluster configurations:

	Base 1x	Base 2x	Base 3x	Base 4x
Process Instances/s	7	12	23	27
Tasks/s	125	217	414	486
Average Backpressure	2%	2%	3%	6%
Write-to-Import Latency	90s	120s	150s	390s
Write-to-Process Latency	140ms	89ms	200ms	160ms
Records Processed Rate	2500	4700	7800	11400
Records Exported Rate	2100	3900	6500	9200

This first observations is that the performance scales particularly well by just adding more resources to the cluster, particularly for a linear increase of the resources the performance as measured by tasks completed increases slightly less than linearly (comparing the 1x and 4x task/s we get 388% the initial rate).

This a very good result as it means that we can scale our system linearly (at least initially) to handle the expected increase in loads.

Importantly, the backpressure is kept low, and the write-to-import latency only increases significantly if we leave the cluster running at max rate for long periods of time. For slightly lower rates the write-to-import latency is kept in the single digits of seconds or lower tens. This might imply that a these sustained max rates, the amount records generated starts to be too much for either ElasticSearch or our web apps that import these records to handle. Some further investigation could be done here to investigate the bottleneck.

Another metric also relevant but not shown in this table is the backlog of records not exported, which kept at almost null through all the experiments conducted.

Bugs found​

During the initial tests, we had several OOM errors in the gateways pods. After some investigation, we found that this was exclusive to the Camunda 8. 6.0 version, which consumes more memory in the gateway than the previous versions. This explains why the gateway memory limits were the only resource that was increased in the new reduced cluster configuration.

In our last Chaos Day we experimented with Operate and different load (Zeebe throughput). We observed that a higher load caused a lower import latency in Operate. The conclusion was that it might be related to Zeebe's exporting configuration, which is affected by a higher load.

In today's chaos day we want to verify how different export and import configurations can affect the importing latency.

TL;DR; We were able to decrease the import latency by ~35% (from 5.7 to 3.7 seconds), by simply reducing the bulk.delay configuration. This worked on low load and even higher load, without significant issues.

🎉 Happy to announce that we are broadening the scope of our Chaos days, to look holistically at the whole Camunda Platform, starting today. In the past Chaos days we often had a close look (or concentrated mostly) at Zeebe performance and stability.

Today, we will look at the Operate import performance and how Zeebe processing throughput might affect (or not?) the throughput and latency of the Operate import. Is it decoupled as we thought?

The import time is an important metric, representing the time until data from Zeebe processing is visible to the User (excluding Elasticsearch's indexing). It is measured from when the record is written to the log, by the Zeebe processor, until Operate reads/imports it from Elasticsearch and converts it into its data model. We got much feedback (and experienced this on our own) that Operate is often lagging behind or is too slow, and of course we want to tackle and investigate this further.

The results from this Chaos day and related benchmarks should allow us to better understand how the current importing of Operate performs, and what its affects. Likely it will be a series of posts to investigate this further. In general, the data will give us some guidance and comparable numbers for the future to improve the importing time. See also related GitHub issue #16912 which targets to improve such.

TL;DR; We were not able to show that Zeebe throughput doesn't affect Operate importing time. We have seen that Operate can be positively affected by the throughput of Zeebe. Surprisingly, Operate was faster to import if Zeebe produced more data (with a higher throughput). One explanation of this might be that Operate was then less idle.

Zeebe 8.6 introduces a new unified flow control mechanism that is able to limit user commands (by default it tries to achieve 200ms response times) and rate limit writes of new records in general (disabled by default). Limiting the write rate is a new feature that can be used to prevent building up an excessive exporting backlog. There are two ways to limit the write rate, either by setting a static limit or by enabling throttling that dynamically adjust the write rate based on the exporting backlog and rate. In these experiments, we will test both ways of limiting the write rate and observe the effects on processing and exporting.

TL;DR; Both setting a static write rate limit and enabling throttling of the write rate can be used to prevent building up an excessive exporting backlog. For users, this will be seen as backpressure because processing speed is limited by the rate at which it can write processing results.

Zeebe 8.6 introduces a new unified flow control mechanism that is able to limit user commands (by default it tries to achieve 200ms response times) and rate limit writes of new records in general (disabled by default).

Limiting the write rate is a new feature that can be used to prevent building up an excessive exporting backlog.

In these experiments we will test what happens with the deployment of endless loops that result in high processing load, and how we can use the new flow control to keep the cluster stable.

TL;DR;

Enabling the write rate limiting can help mitigate the effects caused by process instances that contain uncontrolled loops by preventing building up an excessive exporting backlog.

With the addition of end-to-end job streaming capabilities in Zeebe, we wanted to measure the improvements in job activation latency:

• How much is a single job activation latency reduced?
• How much is the activation latency reduced between each task of the same process instance?
• How much is the activation latency reduced on large clusters with a high broker and partition count?

Additionally, we wanted to guarantee that every component involved in streaming, including clients, would remain resilient in the face of load surges.

TL;DR; Job activation latency is greatly reduced, with task based workloads seeing up to 50% reduced overall execution latency. Completing a task now immediately triggers pushing out the next one, meaning the latency to activate the next task in a sequence is bounded by how much time it takes to process its completion in Zeebe. Activation latency is unaffected by how many partitions or brokers there in a cluster, as opposed to job polling, thus ensuring scalability of the system. Finally, reuse of gRPC's flow control mechanism ensure clients cannot be overloaded even in the face of load surges, without impacting other workloads in the cluster.

Head over to the documentation to learn how to start using job push!

With Zeebe now supporting the addition and removal of brokers to a running cluster, we wanted to test three things:

1. Is there an impact on processing performance while scaling?
2. Is scaling resilient to high processing load?
3. Can scaling up improve processing performance?

TL;DR; Scaling up works even under high load and has low impact on processing performance. After scaling is complete, processing performance improves in both throughput and latency.

We continue our previous experiments with dynamically scaling by now also testing whether the cluster survives dataloss during the process.

One goal is to verify that we haven't accidentally introduced a single point of failure in the cluster. Another is to ensure that data loss does not corrupt the cluster topology.

TL;DR; Even with dataloss, the scaling completes successfully and with the expected results. We found that during scaling, a single broker of the previous cluster configuration can become a single point of failure by preventing a partition from electing a leader. This is not exactly a bug, but something that we want to improve.