Enhancing Istio Propagation Delay | by Ying Zhu | The Airbnb Tech Weblog | Mar, 2023

A case examine in service mesh efficiency optimization

On this article, we’ll showcase how we recognized and addressed a service mesh efficiency drawback at Airbnb, offering insights into the method of troubleshooting service mesh points.


At Airbnb, we use a microservices structure, which requires environment friendly communication between companies. Initially, we developed a homegrown service discovery system known as Smartstack precisely for this function. As the corporate grew, nevertheless, we encountered scalability issues¹. To deal with this, in 2019, we invested in a contemporary service mesh answer known as AirMesh, constructed on the open-source Istio software program. At present, over 90% of our manufacturing visitors has been migrated to AirMesh, with plans to finish the migration by 2023.

The Symptom: Elevated Propagation Delay

After we upgraded Istio from 1.11 to 1.12, we observed a puzzling enhance within the propagation delay — the time between when the Istio management aircraft will get notified of a change occasion and when the change is processed and pushed to a workload. This delay is necessary for our service homeowners as a result of they rely on it to make essential routing choices. For instance, servers must have a swish shutdown interval longer than the propagation delay, in any other case purchasers can ship requests to already-shut-down server workloads and get 503 errors.

Knowledge Gathering: Propagation Delay Metrics

Right here’s how we found the situation: we had been monitoring the Istio metric pilot_proxy_convergence_time for propagation delay after we observed a rise from 1.5 seconds (p90 in Istio 1.11) to 4.5 seconds (p90 in Istio 1.12). Pilot_proxy_convergence_time is one among a number of metrics Istio data for propagation delay. The whole listing of metrics is:

  • pilot_proxy_convergence_time — measures the time from when a push request is added to the push queue to when it’s processed and pushed to a workload proxy. (Notice that change occasions are transformed into push requests and are batched by a course of known as debounce earlier than being added to the queue, which we are going to go into particulars later.)
  • pilot_proxy_queue_time — measures the time between a push request enqueue and dequeue.
  • pilot_xds_push_time — measures the time for constructing and sending the xDS sources. Istio leverages Envoy as its information aircraft. Istiod, the management aircraft of Istio, configures Envoy by the xDS API (the place x will be considered as a variable, and DS stands for discovery service).
  • pilot_xds_send_time — measures the time for truly sending the xDS sources.
A excessive degree graph to assist perceive the metrics associated to propagation delay.

xDS Lock Rivalry

CPU profiling confirmed no noticeable adjustments between 1.11 and 1.12, however dealing with push requests took longer, indicating time was spent on some ready occasions. This led to the suspicion of lock rivalry points.

  • Endpoint Discovery Service (EDS) — describes how one can uncover members of an upstream cluster.
  • Cluster Discovery Service (CDS) — describes how one can uncover upstream clusters used throughout routing.
  • Route Discovery Service (RDS) –describes how one can uncover the route configuration for an HTTP connection supervisor filter at runtime.
  • Listener Discovery Service (LDS) –describes how one can uncover the listeners at runtime.
  • Management aircraft:
    – 1 Istiod pod (reminiscence 26 G, cpu 10 cores)
  • Knowledge aircraft:
    – 50 companies and 500 pods
    – We mimicked adjustments by restarting deployments randomly each 10 seconds and altering digital service routings randomly each 5 seconds
A desk of results² for the perfomance testing.


Right here’s a twist in our analysis: throughout the deep dive of Istio code base, we realized that pilot_proxy_convergence_time doesn’t truly totally seize propagation delay. We noticed in our manufacturing that 503 errors occur throughout server deployment even after we set swish shutdown time longer than pilot_proxy_convergence_time. This metric doesn’t precisely mirror what we wish it to mirror and we have to redefine it. Let’s revisit our community diagram, zoomed out to incorporate the debounce course of to seize the complete lifetime of a change occasion.

A excessive degree diagram of the lifetime of a change occasion.
A CPU profile of Istiod.
A CPU profile of Istiod after DeepCopy enchancment.

To conclude our analysis, we discovered that:

  • We must always use each pilot_debounce_time and pilot_proxy_convergence_time to trace propagation delay.
  • xDS cache can assist with CPU utilization however can impression propagation delay as a consequence of lock rivalry, tune PILOT_ENABLE_CDS_CACHE & PILOT_ENABLE_RDS_CACHE to see what’s greatest on your system.
  • Prohibit the visibility of your Istio manifests by setting the exportTo subject.

Because of the Istio group for creating an important open supply undertaking and for collaborating with us to make it even higher. Additionally name out to the entire AirMesh workforce for constructing, sustaining and enhancing the service mesh layer at Airbnb. Because of Lauren Mackevich, Mark Giangreco and Surashree Kulkarni for modifying the submit.