Terraform setup

[KaseyW31]

Sets up Terraform and an alarm. Also makes some changes to the Dockerfile so that tests can run (issues with deprecations). I don't know much about Ruby/this setup so other changes may be needed

Update: the alarm was deleted somehow so I removed the import and updated the terraform plan output

Questions for reviewers:

1. An alarm currently exists with the same name in Cloudwatch. Would it be preferred to delete the original and use this config to define a new one, or modify the existing one and import it?
2. The existing alarm is set on a log filter on the metric SCSBusterErrorAlarm. In this PR, an new metric SCSBusterError and filter (identical except for name) are defined to keep with the current naming convention. Would using the original metric be preferred?
3. Assuming that we currently don't want an alarm for QA - could either keep the current modularity to allow for a QA setup if wanted, or I could condense everything into one module

terraform plan output:

Terraform used the selected providers to generate the following execution plan. Resource actions are indicated with the following symbols:
  + create

Terraform will perform the following actions:

  # aws_cloudwatch_log_metric_filter.log_error will be created
  + resource "aws_cloudwatch_log_metric_filter" "log_error" {
      + apply_on_transformed_logs = (known after apply)
      + id                        = (known after apply)
      + log_group_name            = "/ecs/scsbuster-production-tf"
      + name                      = "SCSBusterLogError"
      + pattern                   = "{ $.level = \"FATAL\" }"
      + region                    = "us-east-1"

      + metric_transformation {
          + name      = "SCSBusterLogError"
          + namespace = "LogMetrics"
          + unit      = "None"
          + value     = "1"
        }
    }

  # aws_cloudwatch_metric_alarm.log_error_alarm will be created
  + resource "aws_cloudwatch_metric_alarm" "log_error_alarm" {
      + actions_enabled                       = true
      + alarm_actions                         = [
          + "arn:aws:sns:us-east-1:946183545209:research-catalog-team-alarms-production",
        ]
      + alarm_description                     = "Triggered when there's 1 or more fatal error log to SCSBuster within 5 minutes."
      + alarm_name                            = "SCSBusterLogErrorAlarm"
      + arn                                   = (known after apply)
      + comparison_operator                   = "GreaterThanOrEqualToThreshold"
      + datapoints_to_alarm                   = 1
      + evaluate_low_sample_count_percentiles = (known after apply)
      + evaluation_periods                    = 1
      + id                                    = (known after apply)
      + metric_name                           = "SCSBusterLogError"
      + namespace                             = "LogMetrics"
      + period                                = 300
      + region                                = "us-east-1"
      + statistic                             = "Sum"
      + tags_all                              = (known after apply)
      + threshold                             = 1
      + treat_missing_data                    = "notBreaching"
    }

Plan: 2 to add, 0 to change, 0 to destroy.

[7emansell]

1. Yes, we want to import the existing alarm and then modify it
2. No, let's go ahead with the correct naming convention (the metric should be named SomethingError and the alarm should be named SomethingErrorAlarm)
3. Yes, let's just keep it on prod for now

[nonword]

Neat. I was not previously aware of import blocks. This seems maybe better than the import command since it documents that this was previously managed outside of terraform and how we imported it (because I always have to look up the right terraform import command).

[KaseyW31]

I just saw that the original alarm (SCSBusterErrorAlarm) was deleted ~3 weeks ago but I have no recollection of doing so... I'll see if I can figure out how it happened but in the meantime, I removed the import block so that a new alarm (with updated name) will be created.

I don't see any fatal error logs in Cloudwatch in the time without alarm coverage, but ideally this PR can be merged in soon to reduce further downtime

[danamansana]

Lock files should be gitignored and not included in commit, otherwise looks good