getredash/redash · error · Exception

Connection refused

Error message

Connection refused

What it means

Raised by the legacy ElasticSearch/Kibana query runner's test_connection when a requests.exceptions.RequestException occurs — i.e. the HTTP request never completed (DNS failure, refused socket, timeout, SSL error). It maps every low-level transport failure to the generic message 'Connection refused', which can be misleading (e.g. a timeout also produces it).

Source

Thrown at redash/query_runner/elasticsearch.py:333

                    add_column_if_needed(mappings, column, column, result_columns, result_columns_index)

                    value = h[column_name][column]
                    row[column] = value[0] if isinstance(value, list) and len(value) == 1 else value

                result_rows.append(row)
        else:
            raise Exception("Redash failed to parse the results it got from Elasticsearch.")

    def test_connection(self):
        try:
            r = requests.get("{0}/_cluster/health".format(self.server_url), auth=self.auth)
            r.raise_for_status()
        except requests.HTTPError as e:
            logger.exception(e)
            raise Exception("Failed to execute query. Return Code: {0}   Reason: {1}".format(r.status_code, r.text))
        except requests.exceptions.RequestException as e:
            logger.exception(e)
            raise Exception("Connection refused")


class Kibana(BaseElasticSearch):
    @classmethod
    def enabled(cls):
        return True

    def _execute_simple_query(self, url, auth, _from, mappings, result_fields, result_columns, result_rows):
        url += "&from={0}".format(_from)
        r = requests.get(url, auth=self.auth)
        r.raise_for_status()

        raw_result = r.json()

        self._parse_results(mappings, result_fields, raw_result, result_columns, result_rows)

        total = raw_result["hits"]["total"]
        result_size = len(raw_result["hits"]["hits"])

View on GitHub (pinned to ca79fe988d)

Solutions

1. Verify the Elasticsearch/Kibana server URL and port (default 9200/5601) in the data source configuration and that the server is running
2. Test reachability from the Redash host: curl -v <url> or telnet host port; check firewall/security groups
3. If the server uses HTTPS or a self-signed cert, correct the URL scheme and certificate settings
4. Prefer the elasticsearch2 query runner (newer) which gives more precise errors

Example fix

// before
url = "http://es.example.com:9200"
// after
url = "https://es.example.com:9200"  # correct scheme/port verified with curl

import socket
url_host, url_port = "es.example.com", 9200
s = socket.socket()
s.settimeout(3)
try:
    s.connect((url_host, url_port))
    reachable = True
except OSError:
    reachable = False
finally:
    s.close()

try:
    runner.test_connection()
except Exception as e:
    if str(e) == "Connection refused":
        # transport-level failure: check URL/port/TLS, not the query
        log_and_alert_network_issue(datasource_url)
    else:
        raise

When it happens

Trigger: Calling test_connection against an Elasticsearch/Kibana URL whose host/port is unreachable, wrong port, firewall blocking the connection, bad hostname, or HTTPS/SSL mismatch. Any RequestException that is not an HTTPError triggers it.

Common situations: Wrong URL/port in the Redash data source settings, Elasticsearch behind a proxy or requiring TLS, cluster down during setup verification, typo in host name.

Understand the failure class

• Connection failures: ECONNREFUSED, ECONNRESET, and friends — why connections get refused, reset, or dropped.

Related errors

• InfluxDB is not healthy. Check logs for more information.
• Username and Password required
• Azure AD Client ID, Client Secret, and Tenant ID are require
• Connection error to: {} {}.
• Couchbase connection error

AI-assisted analysis of getredash/redash@ca79fe988d (2026-08-28). Data as JSON: /api/errors/6446847120d08cfb. Report an issue: GitHub.