fix: add retry logs (#192)

empirical-run · Apr 30, 2024 · 51f2f52 · 51f2f52
1 parent 9f15c2d
commit 51f2f52
Show file tree

Hide file tree

Showing 6 changed files with 26 additions and 7 deletions.
diff --git a/.changeset/lucky-needles-change.md b/.changeset/lucky-needles-change.md
@@ -0,0 +1,7 @@
+---
+"@empiricalrun/ai": patch
+"@empiricalrun/cli": patch
+"@empiricalrun/core": patch
+---
+
+fix: add retry logs for model api calls
diff --git a/packages/ai/src/providers/anthropic/index.ts b/packages/ai/src/providers/anthropic/index.ts
@@ -69,7 +69,7 @@ const createChatCompletion: ICreateChatCompletion = async (body) => {
   try {
     const startedAt = Date.now();
     const response = await promiseRetry<Anthropic.Messages.Message>(
-      (retry) => {
+      (retry, attempt) => {
         return anthropic.messages
           .create({
             model: canonicalModelName(model),
@@ -92,6 +92,9 @@ const createChatCompletion: ICreateChatCompletion = async (body) => {
               err instanceof Anthropic.APIConnectionTimeoutError ||
               err instanceof Anthropic.InternalServerError
             ) {
+              console.warn(
+                `Retrying request for anthropic model: ${body.model}. Retry attempt: ${attempt}`,
+              );
               retry(err);
               throw err;
             }

diff --git a/packages/ai/src/providers/azure-openai/index.ts b/packages/ai/src/providers/azure-openai/index.ts
@@ -39,8 +39,8 @@ const createChatCompletion: ICreateChatCompletion = async (
       maxRetries: 5,
       shouldRetry: async (resp, retryCount) => {
         if (resp instanceof Response) {
-          console.log(
-            `Retrying request for azure-openai model: ${body.model}. Retry count: ${retryCount}`,
+          console.warn(
+            `Retrying request for azure-openai model: ${body.model}. Retry attempt: ${retryCount}`,
           );
           if (resp.status === 429) {
             requestStartTime = new Date().getTime();

diff --git a/packages/ai/src/providers/fireworks/index.ts b/packages/ai/src/providers/fireworks/index.ts
@@ -40,7 +40,7 @@ const createChatCompletion: ICreateChatCompletion = async (body) => {
   try {
     const startedAt = Date.now();
     const completion = await promiseRetry<IChatCompletion>(
-      (retry) => {
+      (retry, attempt) => {
         return fetch("https://api.fireworks.ai/inference/v1/chat/completions", {
           method: "POST",
           headers: {
@@ -64,6 +64,9 @@ const createChatCompletion: ICreateChatCompletion = async (body) => {
               AIErrorEnum.RATE_LIMITED,
               "Fireworks API rate limit reached",
             );
+            console.warn(
+              `Retrying request for fireworks model: ${body.model}. Got response status code ${response.status}. Retry attempt: ${attempt}`,
+            );
             retry(err);
             throw err;
           }

diff --git a/packages/ai/src/providers/google/index.ts b/packages/ai/src/providers/google/index.ts
@@ -90,7 +90,7 @@ const createChatCompletion: ICreateChatCompletion = async (body) => {
   try {
     const startedAt = Date.now();
     const completion = await promiseRetry<GenerateContentResult>(
-      (retry) => {
+      (retry, attempt) => {
         return modelInstance
           .startChat({
             // @ts-ignore same as above
@@ -101,6 +101,9 @@ const createChatCompletion: ICreateChatCompletion = async (body) => {
           .catch((err: Error) => {
             // TODO: Replace with instanceof checks when the Gemini SDK exports errors
             if (err.message.includes("[429 Too Many Requests]")) {
+              console.warn(
+                `Retrying request for google model: ${body.model}. Retry attempt: ${attempt}`,
+              );
               retry(err);
             }
             throw err;

diff --git a/packages/ai/src/providers/openai/index.ts b/packages/ai/src/providers/openai/index.ts
@@ -33,7 +33,7 @@ const createChatCompletion: ICreateChatCompletion = async (body) => {
   try {
     const startedAt = Date.now();
     const completions = await promiseRetry<IChatCompletion>(
-      (retry) => {
+      (retry, attempt) => {
         return openai.chat.completions.create(body).catch((err) => {
           if (
             err instanceof OpenAI.RateLimitError &&
@@ -46,6 +46,9 @@ const createChatCompletion: ICreateChatCompletion = async (body) => {
             err instanceof OpenAI.APIConnectionTimeoutError ||
             err instanceof OpenAI.InternalServerError
           ) {
+            console.warn(
+              `Retrying request for openai model: ${body.model}. Retry attempt: ${attempt}`,
+            );
             retry(err);
             throw err;
           }
@@ -158,7 +161,7 @@ const runAssistant: ICreateAndRunAssistantThread = async (body) => {
           return asstRunResp;
         })().catch((err: any) => {
           if ((err.message as string).includes("server_error")) {
-            console.log(
+            console.warn(
               `Retrying request due to server error (attempt ${attempt})`,
             );
             requestStartTime = Date.now();