CVE-2026-32794: Apache Airflow Databricks Provider Flaw

CVE-2026-32794 Overview

CVE-2026-32794 is an Improper Certificate Validation vulnerability (CWE-295) in the Apache Airflow Provider for Databricks. The provider code failed to validate SSL/TLS certificates when establishing connections to Databricks back-end services, creating a significant security gap that could allow attackers to intercept and manipulate network traffic or exfiltrate credentials without detection.

This vulnerability affects Apache Airflow environments that rely on the Databricks provider for orchestrating data pipelines, potentially exposing sensitive credentials and data to man-in-the-middle (MITM) attacks on network paths between Airflow workers and Databricks services.

Critical Impact

Man-in-the-middle attackers can intercept Databricks API traffic, steal authentication credentials, and manipulate data pipeline communications without detection due to missing certificate validation.

Affected Products

• Apache Airflow Provider for Databricks versions 1.10.0 through 1.11.x
• Apache Airflow deployments using the affected Databricks provider versions
• Data pipeline environments connecting to Databricks services via the vulnerable provider

Discovery Timeline

• 2026-03-30 - CVE-2026-32794 published to NVD
• 2026-04-02 - Last updated in NVD database

Technical Details for CVE-2026-32794

Vulnerability Analysis

This certificate validation vulnerability stems from the Apache Airflow Databricks provider's failure to properly verify SSL/TLS certificates when establishing HTTPS connections to Databricks API endpoints. When certificate validation is disabled or improperly implemented, the client accepts any certificate presented by a server—including those generated by attackers positioned on the network path.

In practical terms, this means that an attacker who can position themselves between an Airflow worker and Databricks services (via DNS spoofing, ARP poisoning, compromised network infrastructure, or rogue access points) can present their own certificate and establish a seemingly secure connection. The vulnerable provider would accept this fraudulent certificate, allowing the attacker to decrypt, inspect, and modify all traffic in transit.

The impact includes potential exposure of Databricks API tokens, workspace credentials, and sensitive data being transferred through pipelines. Organizations running Airflow in cloud environments with shared network infrastructure or over untrusted networks are particularly at risk.

Root Cause

The root cause is the absence of proper SSL/TLS certificate chain validation in the Databricks provider's HTTP client configuration. When making API requests to Databricks endpoints, the provider either explicitly disabled certificate verification or failed to configure proper certificate authority (CA) validation, effectively treating all certificates as trusted regardless of their legitimacy.

This type of flaw typically occurs when developers disable certificate verification during development or testing (to bypass self-signed certificate warnings) and inadvertently leave it disabled in production code.

Attack Vector

Exploitation requires an attacker to achieve a man-in-the-middle position on the network path between the vulnerable Airflow instance and Databricks services. This can be accomplished through:

1. Network-level attacks: ARP spoofing, DNS poisoning, or BGP hijacking to redirect traffic through attacker-controlled infrastructure
2. Compromised network infrastructure: Exploiting routers, switches, or proxies in the traffic path
3. Rogue access points: In environments where Airflow workers connect over wireless networks
4. Cloud network compromise: Exploiting misconfigurations in cloud virtual networks

Once positioned, the attacker intercepts the TLS handshake, presents their own certificate, and establishes separate encrypted sessions with both the Airflow worker and Databricks—relaying and potentially modifying traffic between them. Since the vulnerable provider doesn't validate certificates, it accepts the attacker's certificate without warning.

The attack requires network access (not just internet access), making this a targeted attack vector rather than an opportunistic one. However, the potential for credential theft and data manipulation makes this a significant concern for organizations handling sensitive data through their Airflow pipelines.

Detection Methods for CVE-2026-32794

Indicators of Compromise

• Unexpected SSL/TLS certificate warnings or errors from other clients connecting to Databricks
• Network traffic analysis showing certificate mismatches between expected Databricks certificates and those observed in traffic
• Unusual API activity patterns from Airflow service accounts on Databricks
• Evidence of credential usage from unexpected IP addresses or geographic locations

Detection Strategies

• Monitor for certificate anomalies in network traffic between Airflow workers and Databricks endpoints using network security monitoring tools
• Implement certificate pinning detection at the network perimeter to identify connections accepting unexpected certificates
• Review Databricks audit logs for suspicious API activity or access from unexpected sources
• Deploy network intrusion detection rules to identify potential MITM attack patterns

Monitoring Recommendations

• Enable comprehensive logging for all Airflow Databricks provider connections and review for anomalies
• Implement network traffic analysis to baseline normal certificate fingerprints for Databricks services
• Configure alerts for authentication failures or unusual access patterns in Databricks workspaces
• Monitor for signs of ARP spoofing or DNS poisoning on networks hosting Airflow infrastructure

How to Mitigate CVE-2026-32794

Immediate Actions Required

• Upgrade the Apache Airflow Provider for Databricks to version 1.12.0 or later immediately
• Audit network infrastructure for potential compromise if vulnerable versions were deployed in production
• Rotate Databricks API tokens and credentials that may have been exposed through vulnerable deployments
• Review Databricks audit logs for suspicious activity during the exposure window

Patch Information

Apache has released version 1.12.0 of the Airflow Provider for Databricks which properly implements SSL/TLS certificate validation. Users can upgrade using pip:

pip install apache-airflow-providers-databricks>=1.12.0

For detailed information on the fix, see the GitHub Pull Request #63704 and the Apache Mailing List announcement.

Workarounds

• If immediate upgrade is not possible, implement network-level controls to ensure Airflow workers only communicate with Databricks through trusted, secured network paths
• Deploy network segmentation to isolate Airflow infrastructure and limit potential MITM attack surfaces
• Use VPN tunnels or private connectivity options (such as Azure Private Link or AWS PrivateLink) for Databricks connections to eliminate exposure to untrusted networks
• Implement network-level certificate validation using proxy servers or security appliances that enforce proper TLS validation

# Upgrade command for Apache Airflow Databricks Provider
pip install --upgrade apache-airflow-providers-databricks>=1.12.0

# Verify installed version
pip show apache-airflow-providers-databricks | grep Version