SShortSingh.
Back to feed

Dynatrace completes Arize acquisition, aiming to link AI engineering and operations monitoring

0
·1 views

Dynatrace has finished acquiring Arize, the company behind the Arize AX platform and the open source Phoenix project. The deal is meant to bridge AI engineers, who trace agents and run evaluations, with SRE and platform teams who monitor infrastructure and services. The article explains that AI systems can fail silently, giving confident but wrong answers without any conventional error signals, so they need additional telemetry such as agent trajectories and retrieved context. Arize AX, the Dynatrace platform and Phoenix will remain available separately, with Phoenix staying open source and building on OpenTelemetry and OpenInference. Dynatrace has not yet published its integration roadmap, which it says will follow in the coming months.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Author's Cache Header Config Failed Silently, Serving Default One-Hour Caching

A developer documented caching headers for their static Next.js site on Firebase Hosting, only to find via curl that the live responses did not match the policy described. Every page and text file was being served with Firebase's default one-hour cache lifetime. The cause was a glob pattern matching request paths, while clean URLs meant readers never requested .html files, so the no-cache rule never applied. A separate rule meant to cache hashed assets for a year also caught manually placed files like a headshot and site icon, which could stay stale for returning visitors after replacement. The author notes the lesson of verifying deployed output rather than only the config file.

0
ProgrammingDEV Community ·

Security checks to run before real users sign up to an AI-built app

AI app builders can produce a working app quickly, but a finished-looking app may still expose personal data. A November 2026 study by Crissy Field reportedly found many Lovable projects, among 159,041 examined, where sensitive information could be read without logging in. Because attackers now also use AI to find old bugs, even small apps are likely targets for automated scanners. The article recommends ten checks before launch, including testing two accounts for cross-account data access, searching client-side code for exposed secret keys, and testing forms and actions while logged out. Failures should be fixed before inviting real users.

0
ProgrammingDEV Community ·

Why AI Apps Fail Under Real Traffic and a Lightweight TypeScript Fix

A DEV Community article explains why generic retry logic breaks down when applied to LLM APIs from providers such as OpenAI, Anthropic, and Gemini. It notes that stacking custom retries on top of built-in SDK retries can multiply requests into traffic storms during outages. The piece also points out that billing-related errors like insufficient_quota should trigger immediate key rotation rather than backoff delays, and that mid-stream failures can duplicate streamed text. Finally, it introduces @rohan9024/retrykit, a zero-dependency TypeScript library designed to handle rate limits, key rotation, circuit breaking, stream recovery, and context compaction.

0
ProgrammingDEV Community ·

Developer builds zero-knowledge proof gatekeeper for AI agent payment decisions

A developer built a local decision engine that judges proof-backed payment requests for zkVerify's Agent Games tournament. The system generates Groth16 zero-knowledge proofs of each decision using the official gatekeeper circuit and verifies them against the exported verification key. In an 11-case test harness, every decision was correct and every provable proof verified, with proving and verification taking well under the 60-second arena deadline. The circuit refused to produce proofs for incorrect decisions, but one trap, proof reuse, cannot be enforced in-circuit and must instead be handled by the platform's seen-set. The author stresses that the work ran locally, with no mainnet transactions or scored tournament rounds.