The X-Ray SDKs and daemon reach end of support in February 2027. Compare Encore's built-in tracing, AWS Distro for OpenTelemetry with CloudWatch, Jaeger, Grafana Tempo, Honeycomb and Datadog APM for tracing a backend on AWS.

AWS X-Ray has been the default way to trace requests through Lambda functions and ECS services for years, with the X-Ray SDK wrapped around HTTP clients and AWS SDK calls and a daemon sidecar forwarding segments.
That path now has a date on it: per the support timeline, the X-Ray SDKs and daemon entered maintenance mode on February 25, 2026 and reach end of support on February 25, 2027, with AWS pointing to OpenTelemetry.
Every X-Ray user therefore has to touch their instrumentation anyway, which makes it a natural point to ask where traces should go and how much instrumentation code the services should carry.
Encore brings infrastructure from code to TypeScript and Go backends, and its runtime records API calls, database queries, Pub/Sub messages and HTTP requests as traces without instrumentation code. Per the tracing docs, traces are captured in all environments, including local development, where encore run shows them in the local development dashboard, and in production you view them in the Encore Platform dashboard. Traces include request and response payloads, with fields marked as sensitive redacted, plus structured logs and stack traces. On deploy to your own AWS account Encore runs services on ECS Fargate or EKS and provisions an IAM role for the service; metrics can be sent to Grafana Cloud or Datadog, and container logs go to CloudWatch.
Run in your terminal to get started locally.
// checkout/checkout.ts
import { api } from "encore.dev/api";
import { SQLDatabase } from "encore.dev/storage/sqldb";
import { Topic } from "encore.dev/pubsub";
import { inventory } from "~encore/clients";
const db = new SQLDatabase("checkout", { migrations: "./migrations" });
export const orderPlaced = new Topic<{ orderId: string }>("order-placed", {
deliveryGuarantee: "at-least-once",
});
// One trace covers the service call, the insert and the publish,
// plus every subscriber that handles the message. No tracing code.
export const placeOrder = api(
{ expose: true, method: "POST", path: "/orders" },
async (p: { sku: string; qty: number }): Promise<{ orderId: string }> => {
await inventory.reserve({ sku: p.sku, qty: p.qty });
const row = await db.queryRow<{ id: string }>`
INSERT INTO orders (sku, qty) VALUES (${p.sku}, ${p.qty}) RETURNING id::text`;
await orderPlaced.publish({ orderId: row!.id });
return { orderId: row!.id };
},
);
The limit is scope: Encore traces Encore services, so a Lambda function or a legacy service outside the app needs OpenTelemetry to show up in a trace. Teams with a mixed estate often use Encore's tracing for the services built on it and keep a collector for the rest.
Traces in local development matter more with coding agents: when an agent's change breaks a request, the local trace shows which call or query failed without adding logging first.
Best fit: TypeScript or Go backends that want every request traced from local development to their own AWS or GCP account, with the instrumentation coming from the runtime.
The path AWS recommends: replace the X-Ray SDK with OpenTelemetry instrumentation, through the AWS Distro for OpenTelemetry (ADOT) or upstream SDKs, and replace the daemon with the CloudWatch agent or an OpenTelemetry Collector using the X-Ray exporter. Traces keep appearing in the X-Ray and CloudWatch consoles, so dashboards and alarms survive. The AWS migration guide maps X-Ray SDK concepts to OpenTelemetry ones. It is the least disruptive option, and the instrumentation becomes portable.
Best fit: teams that want to keep traces in CloudWatch and only replace the deprecated pieces.
Jaeger is the CNCF tracing backend, and since version 2 it is built on the OpenTelemetry Collector, so OpenTelemetry data goes in natively. The all-in-one container is a convenient local backend. In production you run the collector and query services plus a storage backend such as Elasticsearch, OpenSearch or Cassandra, and that storage is where the operational work sits.
Best fit: teams that want an open source tracing backend under their control and already run a search or wide-column store.
Tempo stores traces in object storage such as S3, which keeps retention cheap, and is queried through Grafana with TraceQL, next to Prometheus metrics and Loki logs. It can be self-hosted or used through Grafana Cloud. Self-hosting means running several components at scale, so most small teams use the hosted version.
Best fit: teams already on Grafana that want traces next to metrics and logs.
Honeycomb is a hosted observability backend built around high-cardinality queries over events and traces, with OpenTelemetry as the main way in. Datadog APM pairs its own tracers and agent, or OpenTelemetry, with the rest of Datadog's monitoring. Both remove the backend operations entirely, and both bill by volume, so sampling decisions become cost decisions.
Best fit: teams that want a hosted tracing backend and are willing to manage sampling for cost.
If you only want to get off the deprecated SDKs, switch to OpenTelemetry through ADOT and keep CloudWatch. If you want to own the backend, Jaeger or Tempo store the same OpenTelemetry data. If you want a hosted product, Honeycomb or Datadog. If the services are TypeScript or Go, Encore traces them automatically from local development to your AWS account, and OpenTelemetry stays for whatever sits outside the app.
No. The X-Ray service is not being deprecated, but the X-Ray SDKs and the X-Ray daemon entered maintenance mode on February 25, 2026 and reach end of support on February 25, 2027. AWS recommends instrumenting with OpenTelemetry or the AWS Distro for OpenTelemetry and replacing the daemon with the CloudWatch agent or an OpenTelemetry Collector.
If you want to stay on AWS's backend, move instrumentation to OpenTelemetry through AWS Distro for OpenTelemetry and keep viewing traces in CloudWatch. If you want to own the backend, Jaeger or Grafana Tempo store OpenTelemetry traces. Honeycomb and Datadog are hosted options. Encore traces TypeScript and Go services automatically, including API calls, database queries and Pub/Sub messages, locally and in your AWS account, without instrumentation code.
Deploy to your cloud on AWS and Google Cloud
without managing infrastructure
Encore lets developers and agents define application infrastructure directly in code, then automatically provisions it from local development to production in your AWS or GCP account.