fix: refine chat retry status UX (#23651)

Follow-up to #23282. The retry and terminal error callouts had a few UX
oddities:

- Auto-retrying states reused backend error text that said "Please try
again" even while the UI was already retrying on behalf of the user.
- Terminal error states also said "Please try again" with no action the
user could take.
- `startup_timeout` had no specific title or retry copy — it fell
through to the generic "Retrying request" heading.
- The kind pill showed raw enum values like `startup_timeout` and
`rate_limit`.
- Terminal error metadata showed a "Retryable" / "Not retryable" label
that does not help users.
- A separate "Provider anthropic" metadata row duplicated information
already present in the message body.
- The `usage-limit` error kind used a hyphen while every backend kind
uses underscores.

Changes:

**Backend (`chaterror/message.go`)**

- Split message generation into `terminalMessage()` and
`retryMessage()`, replacing the old `userFacingMessage()`.
- Terminal messages include HTTP status codes and actionable guidance
(e.g. "Check the API key, permissions, and billing settings.").
- Retry messages are clean factual statements without status codes or
remediation, suitable for the retry countdown UI (e.g. "Anthropic is
temporarily overloaded.").
- Removed "Please try again" / "Please try again later" from all paths.
- `StreamRetryPayload` calls `retryMessage()` instead of forwarding
`classified.Message`.

**Frontend**

- Removed the parallel frontend message-generation system:
`getRetryMessage()`, `getProviderDisplayName()`,
`getRetryProviderSubject()`, and the `PROVIDER_DISPLAY_NAMES` map are
all deleted from `chatStatusHelpers.ts`.
- `liveStatusModel.ts` passes `retryState.error` through directly — the
backend owns the copy.
- Added specific title and retry copy for `startup_timeout`, and
extended the title mapping to cover `auth` and `config`.
- Kind pills now show humanized labels ("Startup timeout", "Rate limit",
etc.) instead of raw enum strings.
- Removed the redundant "Provider anthropic" metadata row.
- Removed the terminal "Retryable" / "Not retryable" badge.
- Normalized `"usage-limit"` → `"usage_limit"` and added it to
`ChatProviderFailureKind` so all error kinds follow the same underscore
convention and live in one enum.

Refs #23282.

coderd/x/chatd/chatd_internal_test.go

@@ -732,7 +732,7 @@ func TestSubscribeDeliversRetryEventViaPubsubOnce(t *testing.T) {
 	expected := &codersdk.ChatStreamRetry{
 		Attempt:    1,
 		DelayMs:    (1500 * time.Millisecond).Milliseconds(),
-		Error:      "OpenAI is rate limiting requests (HTTP 429). Please try again later.",
+		Error:      "OpenAI is rate limiting requests (HTTP 429).",
 		Kind:       chaterror.KindRateLimit,
 		Provider:   "openai",
 		StatusCode: 429,
@@ -773,7 +773,7 @@ func TestSubscribePrefersStructuredErrorPayloadViaPubsub(t *testing.T) {
 	defer cancel()
 
 	classified := chaterror.ClassifiedError{
-		Message:    "OpenAI is rate limiting requests (HTTP 429). Please try again later.",
+		Message:    "OpenAI is rate limiting requests (HTTP 429).",
 		Kind:       chaterror.KindRateLimit,
 		Provider:   "openai",
 		Retryable:  true,

coderd/x/chatd/chaterror/classify.go

@@ -178,7 +178,7 @@ func normalizeClassification(classified ClassifiedError) ClassifiedError {
 		classified.Kind = KindGeneric
 	}
 	if classified.Message == "" {
-		classified.Message = userFacingMessage(classified)
+		classified.Message = terminalMessage(classified)
 	}
 	return classified
 }

coderd/x/chatd/chaterror/classify_test.go

@@ -22,7 +22,7 @@ func TestClassify(t *testing.T) {
 			name: "AmbiguousOverloadKeepsProviderUnknown",
 			err:  xerrors.New("status 529 from upstream"),
 			want: chaterror.ClassifiedError{
-				Message:    "The AI provider is temporarily overloaded (HTTP 529). Please try again later.",
+				Message:    "The AI provider is temporarily overloaded (HTTP 529).",
 				Kind:       chaterror.KindOverloaded,
 				Provider:   "",
 				Retryable:  true,
@@ -33,7 +33,7 @@ func TestClassify(t *testing.T) {
 			name: "ExplicitAnthropicOverload",
 			err:  xerrors.New("anthropic overloaded_error"),
 			want: chaterror.ClassifiedError{
-				Message:    "Anthropic is temporarily overloaded. Please try again later.",
+				Message:    "Anthropic is temporarily overloaded.",
 				Kind:       chaterror.KindOverloaded,
 				Provider:   "anthropic",
 				Retryable:  true,
@@ -110,7 +110,7 @@ func TestClassify(t *testing.T) {
 			name: "ExplicitStatus429ClassifiesAsRateLimit",
 			err:  xerrors.New("status 429 from upstream"),
 			want: chaterror.ClassifiedError{
-				Message:    "The AI provider is rate limiting requests (HTTP 429). Please try again later.",
+				Message:    "The AI provider is rate limiting requests (HTTP 429).",
 				Kind:       chaterror.KindRateLimit,
 				Provider:   "",
 				Retryable:  true,
@@ -132,7 +132,7 @@ func TestClassify(t *testing.T) {
 			name: "ServiceUnavailableClassifiesAsRetryableTimeout",
 			err:  xerrors.New("service unavailable"),
 			want: chaterror.ClassifiedError{
-				Message:    "The AI provider is temporarily unavailable. Please try again later.",
+				Message:    "The AI provider is temporarily unavailable.",
 				Kind:       chaterror.KindTimeout,
 				Provider:   "",
 				Retryable:  true,
@@ -176,7 +176,7 @@ func TestClassify(t *testing.T) {
 			name: "DeadlineExceededStaysNonRetryableTimeout",
 			err:  context.DeadlineExceeded,
 			want: chaterror.ClassifiedError{
-				Message:    "The request timed out before it completed. Please try again.",
+				Message:    "The request timed out before it completed.",
 				Kind:       chaterror.KindTimeout,
 				Provider:   "",
 				Retryable:  false,
@@ -279,7 +279,7 @@ func TestClassify_TransportFailuresUseBroaderRetryMessage(t *testing.T) {
 			require.True(t, classified.Retryable)
 			require.Equal(
 				t,
-				"The AI provider is temporarily unavailable. Please try again later.",
+				"The AI provider is temporarily unavailable.",
 				classified.Message,
 			)
 		})
@@ -299,7 +299,7 @@ func TestClassify_StartupTimeoutWrappedClassificationWins(t *testing.T) {
 	)
 
 	require.Equal(t, chaterror.ClassifiedError{
-		Message:    "OpenAI did not start responding in time. Please try again.",
+		Message:    "OpenAI did not start responding in time.",
 		Kind:       chaterror.KindStartupTimeout,
 		Provider:   "openai",
 		Retryable:  true,
@@ -315,7 +315,7 @@ func TestWithProviderUsesExplicitHint(t *testing.T) {
 
 	enriched := classified.WithProvider("azure openai")
 	require.Equal(t, chaterror.ClassifiedError{
-		Message:    "Azure OpenAI is rate limiting requests (HTTP 429). Please try again later.",
+		Message:    "Azure OpenAI is rate limiting requests (HTTP 429).",
 		Kind:       chaterror.KindRateLimit,
 		Provider:   "azure",
 		Retryable:  true,
@@ -331,7 +331,7 @@ func TestWithProviderAddsProviderWhenUnknown(t *testing.T) {
 
 	enriched := classified.WithProvider("openai")
 	require.Equal(t, chaterror.ClassifiedError{
-		Message:    "OpenAI is rate limiting requests (HTTP 429). Please try again later.",
+		Message:    "OpenAI is rate limiting requests (HTTP 429).",
 		Kind:       chaterror.KindRateLimit,
 		Provider:   "openai",
 		Retryable:  true,

coderd/x/chatd/chaterror/message.go

@@ -5,78 +5,113 @@ import (
 	"strings"
 )
 
-func userFacingMessage(classified ClassifiedError) string {
+// terminalMessage produces the user-facing error description shown
+// when retries are exhausted. It includes HTTP status codes and
+// actionable remediation guidance.
+func terminalMessage(classified ClassifiedError) string {
 	subject := providerSubject(classified.Provider)
 	switch classified.Kind {
 	case KindOverloaded:
-		return optionalStatusMessage(
-			subject,
-			classified.StatusCode,
-			"%s is temporarily overloaded (HTTP %d). Please try again later.",
-			"%s is temporarily overloaded. Please try again later.",
-		)
+		if classified.StatusCode > 0 {
+			return fmt.Sprintf(
+				"%s is temporarily overloaded (HTTP %d).",
+				subject, classified.StatusCode,
+			)
+		}
+		return fmt.Sprintf("%s is temporarily overloaded.", subject)
+
 	case KindRateLimit:
-		return optionalStatusMessage(
-			subject,
-			classified.StatusCode,
-			"%s is rate limiting requests (HTTP %d). Please try again later.",
-			"%s is rate limiting requests. Please try again later.",
-		)
+		if classified.StatusCode > 0 {
+			return fmt.Sprintf(
+				"%s is rate limiting requests (HTTP %d).",
+				subject, classified.StatusCode,
+			)
+		}
+		return fmt.Sprintf("%s is rate limiting requests.", subject)
+
 	case KindTimeout:
 		if classified.StatusCode > 0 {
 			return fmt.Sprintf(
-				"%s is temporarily unavailable (HTTP %d). Please try again later.",
-				subject,
-				classified.StatusCode,
+				"%s is temporarily unavailable (HTTP %d).",
+				subject, classified.StatusCode,
 			)
 		}
-		if classified.Retryable {
-			return fmt.Sprintf("%s is temporarily unavailable. Please try again later.", subject)
+		if !classified.Retryable {
+			return "The request timed out before it completed."
 		}
-		return "The request timed out before it completed. Please try again."
+		return fmt.Sprintf("%s is temporarily unavailable.", subject)
+
 	case KindStartupTimeout:
-		return fmt.Sprintf("%s did not start responding in time. Please try again.", subject)
+		return fmt.Sprintf(
+			"%s did not start responding in time.", subject,
+		)
+
 	case KindAuth:
-		if displayName := providerDisplayName(classified.Provider); displayName != "" {
-			return fmt.Sprintf(
-				"Authentication with %s failed. Check the API key, permissions, and billing settings.",
-				displayName,
-			)
+		displayName := providerDisplayName(classified.Provider)
+		if displayName == "" {
+			displayName = "the AI provider"
 		}
-		return "Authentication with the AI provider failed. Check the API key, permissions, and billing settings."
+		return fmt.Sprintf(
+			"Authentication with %s failed."+
+				" Check the API key, permissions, and billing settings.",
+			displayName,
+		)
+
 	case KindConfig:
 		return fmt.Sprintf(
-			"%s rejected the model configuration. Check the selected model and provider settings.",
+			"%s rejected the model configuration."+
+				" Check the selected model and provider settings.",
 			subject,
 		)
+
 	default:
 		if classified.StatusCode > 0 {
-			suffix := " Please try again."
-			if classified.Retryable {
-				suffix = " Please try again later."
-			}
 			return fmt.Sprintf(
-				"%s returned an unexpected error (HTTP %d).%s",
-				subject,
-				classified.StatusCode,
-				suffix,
+				"%s returned an unexpected error (HTTP %d).",
+				subject, classified.StatusCode,
 			)
 		}
-		if classified.Retryable {
-			return fmt.Sprintf(
-				"%s returned an unexpected error. Please try again later.",
-				subject,
-			)
+		if !classified.Retryable {
+			return "The chat request failed unexpectedly."
 		}
-		return "The chat request failed unexpectedly. Please try again."
+		return fmt.Sprintf("%s returned an unexpected error.", subject)
 	}
 }
 
-func optionalStatusMessage(subject string, statusCode int, withStatus string, withoutStatus string) string {
-	if statusCode > 0 {
-		return fmt.Sprintf(withStatus, subject, statusCode)
+// retryMessage produces a clean factual description suitable for
+// display alongside the retry countdown UI. It omits HTTP status
+// codes (surfaced separately in the payload) and remediation
+// guidance (not actionable while auto-retrying).
+func retryMessage(classified ClassifiedError) string {
+	subject := providerSubject(classified.Provider)
+	switch classified.Kind {
+	case KindOverloaded:
+		return fmt.Sprintf("%s is temporarily overloaded.", subject)
+	case KindRateLimit:
+		return fmt.Sprintf("%s is rate limiting requests.", subject)
+	case KindTimeout:
+		return fmt.Sprintf("%s is temporarily unavailable.", subject)
+	case KindStartupTimeout:
+		return fmt.Sprintf(
+			"%s did not start responding in time.", subject,
+		)
+	case KindAuth:
+		displayName := providerDisplayName(classified.Provider)
+		if displayName == "" {
+			displayName = "the AI provider"
+		}
+		return fmt.Sprintf(
+			"Authentication with %s failed.", displayName,
+		)
+	case KindConfig:
+		return fmt.Sprintf(
+			"%s rejected the model configuration.", subject,
+		)
+	default:
+		return fmt.Sprintf(
+			"%s returned an unexpected error.", subject,
+		)
 	}
-	return fmt.Sprintf(withoutStatus, subject)
 }
 
 func providerSubject(provider string) string {

coderd/x/chatd/chaterror/payload.go

@@ -30,7 +30,7 @@ func StreamRetryPayload(
 	return &codersdk.ChatStreamRetry{
 		Attempt:    attempt,
 		DelayMs:    delay.Milliseconds(),
-		Error:      classified.Message,
+		Error:      retryMessage(classified),
 		Kind:       classified.Kind,
 		Provider:   classified.Provider,
 		StatusCode: classified.StatusCode,

coderd/x/chatd/chaterror/payload_test.go

@@ -20,7 +20,7 @@ func TestStreamErrorPayloadUsesNormalizedClassification(t *testing.T) {
 	payload := chaterror.StreamErrorPayload(classified)
 
 	require.Equal(t, &codersdk.ChatStreamError{
-		Message:    "Azure OpenAI is rate limiting requests (HTTP 429). Please try again later.",
+		Message:    "Azure OpenAI is rate limiting requests (HTTP 429).",
 		Kind:       chaterror.KindRateLimit,
 		Provider:   "azure",
 		Retryable:  true,
@@ -40,7 +40,7 @@ func TestStreamRetryPayloadUsesNormalizedClassification(t *testing.T) {
 	delay := 3 * time.Second
 	startedAt := time.Now()
 	payload := chaterror.StreamRetryPayload(2, delay, chaterror.ClassifiedError{
-		Message:    "retry me",
+		Message:    "OpenAI returned an unexpected error (HTTP 503).",
 		Kind:       chaterror.KindGeneric,
 		Provider:   "openai",
 		Retryable:  true,
@@ -50,7 +50,9 @@ func TestStreamRetryPayloadUsesNormalizedClassification(t *testing.T) {
 	require.NotNil(t, payload)
 	require.Equal(t, 2, payload.Attempt)
 	require.Equal(t, delay.Milliseconds(), payload.DelayMs)
-	require.Equal(t, "retry me", payload.Error)
+	// Retry messages omit the HTTP status code; the status code is
+	// surfaced separately in the payload's StatusCode field.
+	require.Equal(t, "OpenAI returned an unexpected error.", payload.Error)
 	require.Equal(t, chaterror.KindGeneric, payload.Kind)
 	require.Equal(t, "openai", payload.Provider)
 	require.Equal(t, 503, payload.StatusCode)

coderd/x/chatd/chatloop/chatloop_test.go

@@ -144,7 +144,7 @@ func TestRun_OnRetryEnrichesProvider(t *testing.T) {
 	require.Equal(t, 429, records[0].classified.StatusCode)
 	require.Equal(
 		t,
-		"OpenAI is rate limiting requests (HTTP 429). Please try again later.",
+		"OpenAI is rate limiting requests (HTTP 429).",
 		records[0].classified.Message,
 	)
 }
@@ -254,7 +254,7 @@ func TestRun_RetriesStartupTimeoutWhileOpeningStream(t *testing.T) {
 	require.Equal(t, "openai", retries[0].Provider)
 	require.Equal(
 		t,
-		"OpenAI did not start responding in time. Please try again.",
+		"OpenAI did not start responding in time.",
 		retries[0].Message,
 	)
 	require.ErrorIs(t, <-attemptCause, errStartupTimeout)
@@ -313,7 +313,7 @@ func TestRun_RetriesStartupTimeoutBeforeFirstPart(t *testing.T) {
 	require.Equal(t, "openai", retries[0].Provider)
 	require.Equal(
 		t,
-		"OpenAI did not start responding in time. Please try again.",
+		"OpenAI did not start responding in time.",
 		retries[0].Message,
 	)
 	require.ErrorIs(t, <-attemptCause, errStartupTimeout)
@@ -475,7 +475,7 @@ func TestRun_RetriesStartupTimeoutWhenStreamClosesSilently(t *testing.T) {
 	require.Equal(t, "openai", retries[0].Provider)
 	require.Equal(
 		t,
-		"OpenAI did not start responding in time. Please try again.",
+		"OpenAI did not start responding in time.",
 		retries[0].Message,
 	)
 	require.ErrorIs(t, <-attemptCause, errStartupTimeout)