> ## Documentation Index > Fetch the complete documentation index at: https://docs.bronto.io/llms.txt > Use this file to discover all available pages before exploring further. # Troubleshooting > Fixes for the errors people hit most often with the bronto CLI, from rejected keys to missing datasets. Every error has a stable code in parentheses (or in the `code` field of the JSON envelope when stderr isn't a terminal) and usually a hint with the fix. These three commands answer most questions: ```sh theme={"dark"} bronto auth status # which profile, key source, region, base URL, and whether the key works bronto ping # can the CLI reach the API, and how fast bronto config list # every resolved setting and where it came from ``` ## Authentication ### `auth_invalid_key` (exit 3) The API rejected the key (HTTP 401). It may be mistyped, revoked, or from another region. Run `bronto auth login` again; it tries both regions and stores the one that works. Check that `BRONTO_API_KEY` isn't set to an old value, because it takes priority over the keychain. ### `auth_insufficient_role` (exit 3) The key is valid but lacks permission (HTTP 403). The usual cause is an ingestion-only key, which can run `bronto send` and nothing else. Create a key with the Search or Admin role ([API Keys](https://docs.bronto.io/Account-Management/API-Keys)) and log in with it. ### `auth_missing_key` (exit 3) No key was found in `--api-key`, `BRONTO_API_KEY`, or the keychain for the active profile. If you use profiles, check that you're on the one you logged into: `bronto auth status` shows the active profile. ### No keychain available In containers, CI, and minimal Linux installs, `bronto auth login` falls back to a `credentials` file with `0600` permissions and prints a warning. To avoid both, set `BRONTO_API_KEY` directly; the keychain isn't touched once a key is resolved. [CI/CD](/cli/guides/ci-cd) has the details. ### "could not read stored credentials" The keychain or credentials file returned an error, for example a corrupt file. The command still runs with any key from the flag or environment. `bronto auth status` shows the underlying error. `bronto auth logout` followed by `bronto auth login` rewrites the entry. ## Datasets and queries ### `usage_missing_dataset` (exit 2) The account has several datasets and the command doesn't know which one to use. The error lists them. Pass `-d `, or set a default: ```sh theme={"dark"} bronto config set default_dataset prod/checkout-service ``` ### `usage_ambiguous_dataset` (exit 2) The name exists in more than one collection. Qualify it: `-d prod/checkout-service` instead of `-d checkout-service`. ### `dataset_not_found` (exit 4) No dataset has that name or id. `bronto datasets list` shows what exists. Names are exact and include the collection when you use the `collection/name` form. ### `usage_invalid_query` (exit 2) The WHERE expression didn't parse. The hint shows a caret under the problem. Check it locally with `bronto query check ""`, and see the [search syntax](https://docs.bronto.io/query-syntax/search-syntax). ### `usage_unknown_field` or `usage_ambiguous_field` (exit 2) A filter flag such as `--eq` named a field the dataset doesn't have, or one that matches several. The error lists candidates. `bronto fields -d ` finds the exact spelling. ### No results, but you expected some * Widen the window. Without a time flag, `search` covers the last 15 minutes. Pass `--since 24h` or an explicit `--from`/`--to`. * Check the dataset with `-d`. Without it, the CLI may have used `default_dataset`. * Print the compiled filter with `--show-query` when you use filter flags. ## Scripts and CI ### `usage_confirmation_required` (exit 2) A destructive command ran without a terminal to confirm on. Pass `--yes`, or `--dry-run` to see what it would do. ### `usage_invalid_flags`: --jq requires -o json or jsonl At a terminal the default format is a table, which `--jq` can't process. Add `-o json` or `-o jsonl`. Piped output works without it. ### `rate_limited` or `timeout` (exit 5) The API throttled the request or it took too long. A rate-limited request, or any read, is safe to retry. A timed-out write (`create`, `update`, `delete`, `send`, a `bronto api` POST) may already have been applied on the server, so check its result, for example with `list` or `get`, before running it again. Idempotent requests are already retried (twice by default) on 429, 502, 503, and 504. Raise the limits with `--max-retries` and `--timeout`, or the matching `BRONTO_MAX_RETRIES` and `BRONTO_TIMEOUT` variables. ### `usage_repl_requires_tty` (exit 2) `bronto repl` is interactive and won't run with piped input or output. Use `bronto search` in scripts. ## Wrong environment ### Requests go to the wrong region or host `bronto ping` prints the URL it used. Override it with `--region`, `BRONTO_REGION`, or for staging and local instances `--base-url` / `BRONTO_BASE_URL`. A `base_url` in a project `.bronto.toml` is ignored on purpose; see [Project files](/cli/configuration/config-files#project-files). ### `config_invalid_region` (exit 2) A region must be a slug such as `eu` or `us`. To use a custom host, set `base_url` instead. ## See what the CLI is doing * `--debug` prints every request and response on stderr: method, URL, status, latency, and truncated bodies. The API key header never appears, and secret fields in bodies show as `REDACTED`. Bodies can still contain log data, so check a trace before you share it. * `--dry-run` prints mutating requests instead of sending them. ## Get help The bronto CLI is an open source, community-supported project. For problems with the CLI, search or open an issue on [GitHub](https://github.com/bronto-community/bronto-cli/issues) rather than contacting Bronto's support team. Include the output of `bronto version` and, if you can, `--debug` output with anything sensitive removed. For problems with the Bronto platform itself, such as your account, data, or billing, see the [Bronto docs](https://docs.bronto.io). This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.