ComposioHQ/composio · error · ErrorDownloadingFile

Error downloading file: {_sanitize_url_for_logging(self.s3ur

Error message

Error downloading file: {_sanitize_url_for_logging(self.s3url)}. Error: {type(e).__name__}

What it means

The GET to the file's S3 URL during download raised a requests RequestException (connection error, DNS failure, TLS problem, timeout at transport level). The error type name and sanitized URL are reported.

Source

Thrown at python/composio/core/models/_files.py:722

            streamed byte count is authoritative because ``Content-Length`` can
            be absent or dishonest.
        """
        # SEC-316: `self.name` also comes from the (potentially compromised or
        # MITM'd) API response. Collapsed to a bare filename and checked against
        # `root` — not `outdir` — so a name like `output_evil/foo`
        # (sibling-prefix attack) is rejected too.
        try:
            outfile = secure_basename_join(outdir, self.name, root=root)
        except UnsafePathComponentError as e:
            raise ErrorDownloadingFile(str(e)) from e
        try:
            response = safe_get(
                self.s3url,
                stream=True,
                timeout=(_CONNECT_TIMEOUT, _READ_TIMEOUT),
            )
        except requests.exceptions.RequestException as e:
            raise ErrorDownloadingFile(
                "Error downloading file: "
                f"{_sanitize_url_for_logging(self.s3url)}. Error: {type(e).__name__}"
            ) from e
        if response.status_code != 200:
            response.close()
            raise ErrorDownloadingFile(
                f"Error downloading file: {_sanitize_url_for_logging(self.s3url)}"
            )

        # Early abort for a self-declared oversized body. The header is only a
        # hint — `parse_content_length` returns None for anything untrustworthy
        # and the streaming counter below is the authoritative limit.
        content_length = parse_content_length(response.headers.get("Content-Length"))
        if content_length is not None and content_length > max_size:
            response.close()
            raise ResponseTooLargeError(
                f"File size ({content_length} bytes) exceeds maximum allowed "
                f"size ({max_size} bytes)"

View on GitHub (pinned to 64b1b85502)

Solutions

  1. Verify the s3url host is reachable from the runtime (curl -I).
  2. Retry the download — transient errors are common.
  3. Fix proxy/TLS configuration if a middlebox is interfering.
  4. Check DNS resolution for the bucket host.

Example fix

# before
path = model.download()
# after
from tenacity import retry, wait_exponential, stop_after_attempt
@retry(wait=wait_exponential(multiplier=1), stop=stop_after_attempt(4))
def safe_download():
    return model.download()
path = safe_download()
Defensive patterns

Strategy: retry

Validate before calling

import socket
from urllib.parse import urlparse

def s3_reachable(url: str) -> bool:
    host = urlparse(url).hostname or ''
    try:
        socket.create_connection((host, 443), timeout=5).close()
        return True
    except OSError:
        return False

Try / catch

from composio.core.models._files import ErrorDownloadingFile

for attempt in range(3):
    try:
        path = model.download()
        break
    except ErrorDownloadingFile as e:
        if attempt == 2:
            raise
        time.sleep(2 ** attempt)

Prevention

When it happens

Trigger: FileModel.download when the runtime cannot reach the S3/CDN host of model.s3url — network outage, egress firewall, proxy blocking object storage, expired DNS.

Common situations: Restricted CI/container networks blocking *.amazonaws.com or the CDN; transient connectivity blips; misconfigured HTTPS_PROXY; air-gapped environments.

Related errors

AI-assisted analysis of ComposioHQ/composio@64b1b85502 (2026-08-28). Data as JSON: /api/errors/a393c33463034f1d. Report an issue: GitHub.