Problem

In Meta infrastructure systems, engineers often need to summarize large volumes of Scuba-style server logs. Given a list of log lines, return the top 5 most frequent error messages.

A log line is considered an error message only if it starts with the prefix "ERROR: ". The message is the substring after that prefix. Count identical messages together and return up to 5 messages ordered by descending frequency. If two messages have the same frequency, return them in lexicographically ascending order.

Formal Specification

Implement a function that takes a list of strings logs and returns a list of strings.

• Input: logs, a list of log lines
• Output: a list containing up to 5 error messages

Ignore all non-error log lines.

Examples

Example 1 Input: logs = ["INFO: boot", "ERROR: disk full", "ERROR: timeout", "ERROR: disk full"] Output: ["disk full", "timeout"] Explanation: Only lines starting with "ERROR: " are counted. "disk full" appears twice and "timeout" appears once.

Example 2 Input: logs = ["ERROR: b", "ERROR: a", "ERROR: b", "ERROR: a", "ERROR: c"] Output: ["a", "b", "c"] Explanation: "a" and "b" both appear twice, so lexicographic order breaks the tie.

Constraints

• 1 <= len(logs) <= 10^5
• 0 <= len(logs[i]) <= 200
• Log lines are ASCII strings
• Return fewer than 5 messages if fewer than 5 distinct error messages exist